Model loading strategy determination method and device, equipment and storage medium

Abstract

The embodiment of the invention provides a model loading strategy determination method and device, equipment and a storage medium, and relates to the technical field of finance science and technology, the method comprises the following steps: firstly obtaining outbound files corresponding to n1 outbound scenes respectively, then determining outbound models corresponding to n1 outbound scenes respectively, and according to outbound numbers in the n1 outbound files, determining the outbound models corresponding to n1 outbound scenes respectively; determining model estimated throughput corresponding to the n1 outbound models respectively; determining a first loading strategy according to the model estimated throughput corresponding to the n1 outbound models and the process expected throughput capable of being provided by one process; the first loading strategy comprises the number m1 of processes to be generated, n2 target outbound models to be loaded by each process in the m1 processes, and process throughput allocated to the n2 target outbound models by each process. The model loading strategy determined through the method is more reasonable and accurate, and waste of process load resources can be effectively reduced.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of financial technology (Fintech), and in particular to a method, device, equipment and storage medium for determining a model loading strategy. Background Art

[0002] With the development of computer technology, more and more technologies are being applied in the financial field. The traditional financial industry is gradually transforming into financial technology (Fintech). However, due to the security and real-time requirements of the financial industry, higher requirements are also placed on technology.

[0003] The intelligent outbound call robot is a robot product that uses speech recognition (SR) technology and natural language processing (NLP) technology to provide intelligent services to corporate customers. Among them, the intelligent outbound call robot can automatically dial and communicate with corporate customers in order to provide intelligent services to corporate customers.

[0004] In order to provide more accurate and intelligent services, the intelligent outbound call robot can call different outbound call models to communicate with corporate customers for different outbound call scenarios. For example, for marketing scenarios, the intelligent outbound call robot can call marketing models trained with marketing corpus; for collection scenarios, the intelligent outbound call robot can call collection models trained with collection corpus.

[0005] In order to facilitate the provision of outbound call services, the intelligent outbound call robot can load different outbound call models at the same time. The intelligent outbound call robot can load all outbound call models into one process and evenly distribute the load resources of the process to each outbound call model, or it can load one outbound call model into one process. However, the above two model loading methods do not take into account that different outbound call models correspond to different outbound call volumes, that is, for outbound call models with smaller outbound call volumes, this will cause a waste of load resources, and for outbound call models with larger outbound call volumes, this will cause a large delay. How to reasonably load each outbound call model according to the outbound call volume of the outbound call model in order to reduce the waste of process load resources is a technical problem that urgently needs to be solved. Summary of the invention

[0006] The embodiments of the present application provide a model loading strategy determination method, apparatus, device and storage medium for reasonably loading each model to reduce the waste of process load resources.

[0007] On the one hand, an embodiment of the present application provides a method for determining a model loading strategy, the method comprising:

[0008] Obtain outbound call files corresponding to n1 outbound call scenarios, wherein the outbound call files include the outbound call number to be dialed; n1 is an integer greater than 0;

[0009] Determine the outbound call models corresponding to the n1 outbound call scenarios, and determine the model estimated throughput corresponding to the n1 outbound call models according to the outbound call numbers in the n1 outbound call files;

[0010] A first loading strategy is determined based on the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that a process can provide. The first loading strategy includes the number of processes m1 to be generated, the n2 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each of the m1 processes; the expected process throughput is used to identify that the load of the process is in an optimal state; the n1 outbound call models include the n2 target outbound call models; n2 is less than or equal to n1, and m1 and n2 are both integers greater than 0.

[0011] Optionally, determining the model estimated throughputs corresponding to the n1 outbound call models respectively according to the outbound call numbers in the n1 outbound call files includes:

[0012] Determine a current outbound call window and a historical outbound call window corresponding to the current outbound call window;

[0013] From the i-th outbound call file, determine the outbound call number within the current outbound call window; i is an integer between [1, n1];

[0014] Determine the historical outbound call number within the historical outbound call window from the historical outbound call file corresponding to the i-th outbound call file;

[0015] The model estimated throughput corresponding to the i-th outbound call model is determined according to the outbound call number in the current outbound call window and the historical outbound call number in the historical outbound call window.

[0016] Optionally, determining the model estimated throughput corresponding to the i-th outbound call model according to the outbound call number located in the current outbound call window and the historical outbound call number located in the historical outbound call window includes:

[0017] Determine the number of outbound call numbers corresponding to the current outbound call window according to the outbound call numbers in the current outbound call window;

[0018] Determine the historical answering rate and the historical average number of interactions corresponding to the historical outbound call window according to the historical outbound call number in the historical outbound call window;

[0019] The model estimated throughput corresponding to the i-th outbound call model is determined according to the number of outbound call numbers corresponding to the current outbound call window, the historical answering rate corresponding to the historical outbound call window, and the historical average number of interactions.

[0020] Optionally, determining the first loading strategy according to the model estimated throughputs respectively corresponding to the n1 outbound call models and the process expected throughput that can be provided by one process includes:

[0021] Determine the model residual throughputs respectively corresponding to the n1 outbound call models according to the model estimated throughputs respectively corresponding to the n1 outbound call models;

[0022] Repeat the following process until the model residual throughputs corresponding to the n1 outbound call models are 0:

[0023] Identify a process;

[0024] Selecting n2 target outbound call models from the n1 outbound call models, and determining the process throughputs respectively allocated by the one process to the n2 target outbound call models according to the expected process throughputs;

[0025] The process throughputs respectively allocated by the one process to the n2 target outbound call models are used to update the model remaining throughputs respectively corresponding to the n1 outbound call models.

[0026] Optionally, the selecting n2 target outbound call models from the n1 outbound call models, and determining the process throughputs respectively allocated by the one process to the n2 target outbound call models according to the expected process throughputs, and updating the model residual throughputs respectively corresponding to the n1 outbound call models by using the process throughputs respectively allocated by the one process to the n2 target outbound call models, comprises:

[0027] Determine an allocated throughput corresponding to the one process, and initialize the allocated throughput corresponding to the one process;

[0028] Repeat the following steps until the allocated throughput corresponding to the one process is within a process throughput range, where the process throughput range is determined according to the expected throughput of the process:

[0029] According to the sorting rule, the model remaining throughputs corresponding to the n1 outbound call models are sorted to obtain a first sequence;

[0030] Determine a target outbound call model according to a model remaining throughput corresponding to the outbound call model located at the first position in the first sequence;

[0031] Determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model;

[0032] The process throughput allocated by the one process to the target outbound call model is used to update the allocated throughput corresponding to the one process, and the model remaining throughputs respectively corresponding to the n1 outbound call models are updated.

[0033] Optionally, the step of sorting the model residual throughputs respectively corresponding to the n1 outbound call models according to a sorting rule to obtain a first sequence includes:

[0034] Determine the number S of determined outbound call models, where the determined outbound call model is a target outbound call model corresponding to the allocated throughput corresponding to the one process;

[0035] In the case that the number S of the determined outbound call models is an even number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in descending order to obtain a first sequence;

[0036] In the case that the number S of the determined outbound call models is an odd number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in ascending order to obtain a first sequence.

[0037] Optionally, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model includes:

[0038] Determining a model maximum throughput based on the expected throughput of the process;

[0039] Determine the model minimum throughput according to the process expected throughput and the model estimated throughputs corresponding to the n1 outbound call models respectively;

[0040] The process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput.

[0041] Optionally, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput includes:

[0042] In a case where the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than or equal to a first preset value, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput or the model minimum throughput corresponding to the target outbound call model;

[0043] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, the process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model and the model maximum throughput.

[0044] Optionally, when a ratio of a model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than or equal to a first preset value, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput or the model minimum throughput corresponding to the target outbound call model includes:

[0045] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than a first preset value, using the model remaining throughput corresponding to the target outbound call model as the process throughput allocated by the one process to the target outbound call model;

[0046] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is equal to a first preset value, the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0047] Optionally, when a ratio of a model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model and the model maximum throughput includes:

[0048] Determining a minimum throughput of the process based on the expected throughput of the process;

[0049] Determine the throughput to be allocated corresponding to the process according to the minimum throughput of the process and the allocated throughput corresponding to the process;

[0050] In the case where the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, if the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the throughput to be allocated corresponding to the one process, and the model minimum throughput;

[0051] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, the process throughput allocated to the target outbound call model by the one process is determined according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput.

[0052] Optionally, if the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the throughput to be allocated corresponding to the one process, and the model minimum throughput, includes:

[0053] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then the process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model and the model minimum throughput; the process throughput control value is determined according to the process throughput range;

[0054] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the process throughput allocated by the one process to the target outbound call model is determined according to the throughput to be allocated corresponding to the one process and the model minimum throughput.

[0055] Optionally, if the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model and the model minimum throughput, including:

[0056] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is not equal to the model estimated throughput corresponding to the target outbound call model, then the model remaining throughput corresponding to the target outbound call model is used as the process throughput allocated by the one process to the target outbound call model;

[0057] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is equal to the model estimated throughput corresponding to the target outbound call model, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0058] Optionally, if the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, determining the process throughput allocated by the one process to the target outbound call model according to the throughput to be allocated corresponding to the one process and the model minimum throughput, including:

[0059] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the maximum value of the throughput to be allocated corresponding to the one process and the model minimum throughput is selected as the process throughput allocated by the one process to the target outbound call model.

[0060] Optionally, if the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput, including:

[0061] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, then determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model minimum throughput, and the to-be-allocated throughput corresponding to the one process;

[0062] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, then the process throughput allocated to the target outbound call model by the process is determined according to the model minimum throughput, the model maximum throughput, and the throughput to be allocated corresponding to the one process.

[0063] Optionally, if the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model minimum throughput, and the to-be-allocated throughput corresponding to the one process, includes:

[0064] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is less than the to-be-allocated throughput corresponding to the one process, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model;

[0065] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is greater than or equal to the throughput to be allocated corresponding to the one process, then the maximum value between the throughput to be allocated corresponding to the one process and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0066] Optionally, if the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model minimum throughput, the model maximum throughput, and the to-be-allocated throughput corresponding to the one process, includes:

[0067] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then the model maximum throughput is used as the process throughput allocated by the one process to the target outbound call model; the process throughput control value is determined according to the process throughput range;

[0068] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the sum of the throughput to be allocated corresponding to the one process and the process throughput control value is used as the process throughput allocated by the one process to the target outbound call model.

[0069] Optionally, it also includes:

[0070] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum process throughput, the first loading strategy is adjusted to obtain the second loading strategy; the second loading strategy includes the number of processes m2 to be generated, the n3 target outbound call models to be loaded by each of the m2 processes, and the process throughput allocated by each of the m2 processes to the n3 target outbound call models; the minimum process throughput is determined based on the expected throughput of the process, m2 is less than or equal to m1, and n3 is an integer greater than 0.

[0071] Optionally, if the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, adjusting the first loading strategy to obtain the second loading strategy includes:

[0072] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is less than the ratio of the expected throughput of the process to the second preset value, the allocated throughput corresponding to the m1th process is allocated to other processes to obtain the second loading strategy; the other processes are processes other than the m1th process among the m1 processes; the second loading strategy includes the number of processes to be generated (m1)-1, the n3 target outbound call models to be loaded by each of the (m1)-1 processes, and the process throughput allocated by each of the (m1)-1 processes to the n3 target outbound call models;

[0073] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is greater than or equal to the ratio of the expected throughput of the process to the second preset value, the allocated throughputs corresponding to the other processes are allocated to the m1th process to obtain the second loading strategy; the second loading strategy includes the number of processes m1 to be generated, n3 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated by each of the m1 processes to the n3 target outbound call models.

[0074] Optionally, it also includes:

[0075] According to the first loading strategy, m1 processes are generated, and n2 target outbound call models are loaded for each of the m1 processes, and corresponding process throughputs are respectively allocated to the n2 target outbound call models loaded for each of the m1 processes.

[0076] Optionally, it also includes:

[0077] Obtaining an outbound call number in the outbound call file, and generating a random number for the outbound call number;

[0078] Determine a target outbound call model corresponding to the one outbound call number, and determine m3 processes where the target outbound call model corresponding to the one outbound call number is located; m3 is less than or equal to m1;

[0079] According to the random number, a target process is selected from the m3 processes, and the one outbound call number is allocated to the target process, so that the target outbound call model corresponding to the one outbound call number deployed in the target process dials the one outbound call number.

[0080] Optionally, selecting a target process from the m3 processes according to the random number includes:

[0081] Determine the process throughputs allocated by the m3 processes to the target outbound call model corresponding to the one outbound call number;

[0082] Determine the value intervals corresponding to the m3 processes respectively according to the process throughputs respectively allocated by the m3 processes to the target outbound call model corresponding to the one outbound call number;

[0083] A process corresponding to the value interval of the random number is selected as the target process.

[0084] Optionally, it also includes:

[0085] According to the monitoring window, the m1 processes are monitored to determine the actual model throughputs corresponding to the m4 processes and the n4 target outbound call models loaded by each of the m4 processes; the throughput of each of the m4 processes is greater than 0; m4 is less than or equal to m1, and n4 is an integer greater than 0;

[0086] According to the actual model throughputs corresponding to the n4 target outbound call models loaded by each of the m4 processes, the first loading strategy is adjusted to obtain a third loading strategy; the third loading strategy includes the number of processes m4 to be generated, the n4 target outbound call models to be loaded by each of the m4 processes, and the process throughput allocated by each of the m4 processes to the n4 target outbound call models.

[0087] On the one hand, an embodiment of the present application provides a model loading strategy determination device, the device comprising:

[0088] An acquisition module, used to acquire outbound call files corresponding to n1 outbound call scenarios, wherein the outbound call files include an outbound call number to be dialed; n1 is an integer greater than 0;

[0089] A determination module, used to determine the outbound call models corresponding to n1 outbound call scenarios respectively, and determine the model estimated throughput corresponding to the n1 outbound call models respectively according to the outbound call numbers in the n1 outbound call files;

[0090] The determination module is further used to determine a first loading strategy based on the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that a process can provide, wherein the first loading strategy includes the number of processes m1 to be generated, the n2 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each of the m1 processes; the expected process throughput is used to identify that the load of the process is in an optimal state; the n1 outbound call models include the n2 target outbound call models; n2 is less than or equal to n1, and m1 and n2 are both integers greater than 0.

[0091] Optionally, the determining module is specifically configured to:

[0092] Determine a current outbound call window and a historical outbound call window corresponding to the current outbound call window;

[0093] From the i-th outbound call file, determine the outbound call number within the current outbound call window; i is an integer between [1, n1];

[0094] Determine the historical outbound call number within the historical outbound call window from the historical outbound call file corresponding to the i-th outbound call file;

[0095] The model estimated throughput corresponding to the i-th outbound call model is determined according to the outbound call number in the current outbound call window and the historical outbound call number in the historical outbound call window.

[0096] Optionally, the determining module is specifically configured to:

[0097] Determine the number of outbound call numbers corresponding to the current outbound call window according to the outbound call numbers in the current outbound call window;

[0098] Determine the historical answering rate and the historical average number of interactions corresponding to the historical outbound call window according to the historical outbound call number in the historical outbound call window;

[0099] The model estimated throughput corresponding to the i-th outbound call model is determined according to the number of outbound call numbers corresponding to the current outbound call window, the historical answering rate corresponding to the historical outbound call window, and the historical average number of interactions.

[0100] Optionally, the determining module is specifically configured to:

[0101] Determine the model residual throughputs respectively corresponding to the n1 outbound call models according to the model estimated throughputs respectively corresponding to the n1 outbound call models;

[0102] Repeat the following process until the model residual throughputs corresponding to the n1 outbound call models are 0:

[0103] Identify a process;

[0104] Selecting n2 target outbound call models from the n1 outbound call models, and determining the process throughputs respectively allocated by the one process to the n2 target outbound call models according to the expected process throughputs;

[0105] The process throughputs respectively allocated by the one process to the n2 target outbound call models are used to update the model remaining throughputs respectively corresponding to the n1 outbound call models.

[0106] Optionally, the determining module is specifically configured to:

[0107] Determine an allocated throughput corresponding to the one process, and initialize the allocated throughput corresponding to the one process;

[0108] Repeat the following steps until the allocated throughput corresponding to the one process is within a process throughput range, where the process throughput range is determined according to the expected throughput of the process:

[0109] According to the sorting rule, the model remaining throughputs corresponding to the n1 outbound call models are sorted to obtain a first sequence;

[0110] Determine a target outbound call model according to a model remaining throughput corresponding to the outbound call model located at the first position in the first sequence;

[0111] Determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model;

[0112] The process throughput allocated by the one process to the target outbound call model is used to update the allocated throughput corresponding to the one process, and the model remaining throughputs respectively corresponding to the n1 outbound call models are updated.

[0113] Optionally, the determining module is specifically configured to:

[0114] Determine the number S of determined outbound call models, where the determined outbound call model is a target outbound call model corresponding to the allocated throughput corresponding to the one process;

[0115] In the case that the number S of the determined outbound call models is an even number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in descending order to obtain a first sequence;

[0116] In the case that the number S of the determined outbound call models is an odd number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in ascending order to obtain a first sequence.

[0117] Optionally, the determining module is specifically configured to:

[0118] Determining a model maximum throughput based on the expected throughput of the process;

[0119] Determine the model minimum throughput according to the process expected throughput and the model estimated throughputs corresponding to the n1 outbound call models respectively;

[0120] The process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput.

[0121] Optionally, the determining module is specifically configured to:

[0122] In a case where the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than or equal to a first preset value, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput or the model minimum throughput corresponding to the target outbound call model;

[0123] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, the process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model and the model maximum throughput.

[0124] Optionally, the determining module is specifically configured to:

[0125] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than a first preset value, using the model remaining throughput corresponding to the target outbound call model as the process throughput allocated by the one process to the target outbound call model;

[0126] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is equal to a first preset value, the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0127] Optionally, the determining module is specifically configured to:

[0128] Determining a minimum throughput of the process based on the expected throughput of the process;

[0129] Determine the throughput to be allocated corresponding to the process according to the minimum throughput of the process and the allocated throughput corresponding to the process;

[0130] In the case where the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, if the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the throughput to be allocated corresponding to the one process, and the model minimum throughput;

[0131] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, the process throughput allocated to the target outbound call model by the one process is determined according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput.

[0132] Optionally, the determining module is specifically configured to:

[0133] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then the process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model and the model minimum throughput; the process throughput control value is determined according to the process throughput range;

[0134] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the process throughput allocated by the one process to the target outbound call model is determined according to the throughput to be allocated corresponding to the one process and the model minimum throughput.

[0135] Optionally, the determining module is specifically configured to:

[0136] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is not equal to the model estimated throughput corresponding to the target outbound call model, then the model remaining throughput corresponding to the target outbound call model is used as the process throughput allocated by the one process to the target outbound call model;

[0137] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is equal to the model estimated throughput corresponding to the target outbound call model, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0138] Optionally, the determining module is specifically configured to:

[0139] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the maximum value of the throughput to be allocated corresponding to the one process and the model minimum throughput is selected as the process throughput allocated by the one process to the target outbound call model.

[0140] Optionally, the determining module is specifically configured to:

[0141] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, then determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model minimum throughput, and the to-be-allocated throughput corresponding to the one process;

[0142] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, then the process throughput allocated to the target outbound call model by the process is determined according to the model minimum throughput, the model maximum throughput, and the throughput to be allocated corresponding to the one process.

[0143] Optionally, the determining module is specifically configured to:

[0144] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is less than the to-be-allocated throughput corresponding to the one process, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model;

[0145] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is greater than or equal to the throughput to be allocated corresponding to the one process, then the maximum value between the throughput to be allocated corresponding to the one process and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0146] Optionally, the determining module is specifically configured to:

[0147] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then the model maximum throughput is used as the process throughput allocated by the one process to the target outbound call model; the process throughput control value is determined according to the process throughput range;

[0148] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the sum of the throughput to be allocated corresponding to the one process and the process throughput control value is used as the process throughput allocated by the one process to the target outbound call model.

[0149] Optionally, an adjustment module is further included, and the adjustment module is specifically used to:

[0150] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum process throughput, the first loading strategy is adjusted to obtain the second loading strategy; the second loading strategy includes the number of processes m2 to be generated, the n3 target outbound call models to be loaded by each of the m2 processes, and the process throughput allocated by each of the m2 processes to the n3 target outbound call models; the minimum process throughput is determined based on the expected throughput of the process, m2 is less than or equal to m1, and n3 is an integer greater than 0.

[0151] Optionally, the adjustment module is specifically used to:

[0152] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is less than the ratio of the expected throughput of the process to the second preset value, the allocated throughput corresponding to the m1th process is allocated to other processes to obtain the second loading strategy; the other processes are processes other than the m1th process among the m1 processes; the second loading strategy includes the number of processes to be generated (m1)-1, the n3 target outbound call models to be loaded by each of the (m1)-1 processes, and the process throughput allocated by each of the (m1)-1 processes to the n3 target outbound call models;

[0153] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is greater than or equal to the ratio of the expected throughput of the process to the second preset value, the allocated throughputs corresponding to the other processes are allocated to the m1th process to obtain the second loading strategy; the second loading strategy includes the number of processes m1 to be generated, n3 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated by each of the m1 processes to the n3 target outbound call models.

[0154] Optionally, an execution module is further included, and the execution module is specifically used to:

[0155] According to the first loading strategy, m1 processes are generated, and n2 target outbound call models are loaded for each of the m1 processes, and corresponding process throughputs are respectively allocated to the n2 target outbound call models loaded for each of the m1 processes.

[0156] Optionally, the execution module is further used for:

[0157] Obtaining an outbound call number in the outbound call file, and generating a random number for the outbound call number;

[0158] Determine a target outbound call model corresponding to the one outbound call number, and determine m3 processes where the target outbound call model corresponding to the one outbound call number is located; m3 is less than or equal to m1;

[0159] According to the random number, a target process is selected from the m3 processes, and the one outbound call number is allocated to the target process, so that the target outbound call model corresponding to the one outbound call number deployed in the target process dials the one outbound call number.

[0160] Optionally, the execution module is specifically used to:

[0161] Determine the process throughputs allocated by the m3 processes to the target outbound call model corresponding to the one outbound call number;

[0162] Determine the value intervals corresponding to the m3 processes respectively according to the process throughputs respectively allocated by the m3 processes to the target outbound call model corresponding to the one outbound call number;

[0163] A process corresponding to the value interval of the random number is selected as the target process.

[0164] Optionally, the adjustment module is further used for:

[0165] According to the monitoring window, the m1 processes are monitored to determine the actual model throughputs corresponding to the m4 processes and the n4 target outbound call models loaded by each of the m4 processes; the throughput of each of the m4 processes is greater than 0; m4 is less than or equal to m1, and n4 is an integer greater than 0;

[0166] According to the actual model throughputs corresponding to the n4 target outbound call models loaded by each of the m4 processes, the first loading strategy is adjusted to obtain a third loading strategy; the third loading strategy includes the number of processes m4 to be generated, the n4 target outbound call models to be loaded by each of the m4 processes, and the process throughput allocated by each of the m4 processes to the n4 target outbound call models.

[0167] On the one hand, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned model loading strategy determination method when executing the program.

[0168] On the one hand, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program executable by a computer device. When the program runs on the computer device, the computer device executes the steps of the above-mentioned model loading strategy determination method.

[0169] In an embodiment of the present application, first obtain the outbound call files corresponding to n1 outbound call scenarios, wherein the outbound call files include the outbound call numbers to be dialed; then determine the outbound call models corresponding to the n1 outbound call scenarios, and determine the model estimated throughputs corresponding to the n1 outbound call models according to the outbound call numbers in the n1 outbound call files; then, determine the first loading strategy according to the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that a process can provide, wherein the first loading strategy includes the number of processes m1 to be generated, n2 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each of the m1 processes. The above model loading process does not directly load all outbound call models into one process, nor does it directly load one outbound call model into one process. Instead, it uses the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that one process can provide to accurately determine the number of processes m1 required to load the above n1 outbound call models, the target outbound call model to be loaded by each of the m1 processes, and the process throughput allocated to each target outbound call model by each of the m1 processes. The model loading strategy determined by the above method fully considers the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that one process can provide. The determined model loading strategy is more reasonable and accurate, which can effectively reduce the waste of process load resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0170] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0171] Figure 1 A schematic diagram of a system architecture provided for an embodiment of the present application;

[0172] Figure 2 A flow chart of a method for determining a model loading strategy provided in an embodiment of the present application;

[0173] Figure 3 A flow chart of a method for determining a model estimated throughput corresponding to an i-th outbound call model provided in an embodiment of the present application;

[0174] Figure 4 A flow chart of a method for determining a model estimated throughput corresponding to an i-th outbound call model provided in an embodiment of the present application;

[0175] Figure 5A schematic diagram of a flow chart of a method for determining a first loading strategy provided in an embodiment of the present application;

[0176] Figure 6 A schematic diagram of a flow chart of a method for determining a first loading strategy provided in an embodiment of the present application;

[0177] Fig. 7A A schematic diagram of a process for obtaining a first sequence provided in an embodiment of the present application;

[0178] Figure 7B A schematic diagram of a process for obtaining a first sequence provided in an embodiment of the present application;

[0179] Figure 8 A flowchart of a method for determining a process throughput allocated to a target outbound call model provided in an embodiment of the present application;

[0180] Fig. 9 A flowchart of a method for updating the allocated throughput corresponding to a process provided in an embodiment of the present application;

[0181] Fig.10 A flowchart of a method for dialing an outbound number provided in an embodiment of the present application;

[0182] Fig.11 A flowchart of a method for determining a target process provided in an embodiment of the present application;

[0183] Fig.12 A flowchart of a method for determining a value interval corresponding to a process provided in an embodiment of the present application;

[0184] Fig.13 A schematic diagram of a flow chart of a method for adjusting a first loading strategy provided in an embodiment of the present application;

[0185] Fig.14 A schematic diagram of a flow chart of a method for adjusting a first loading strategy provided in an embodiment of the present application;

[0186] Fig.15 A schematic diagram of a flow chart of a method for adjusting a first loading strategy provided in an embodiment of the present application;

[0187] Fig.16 A schematic diagram of a flow chart of a method for adjusting a first loading strategy provided in an embodiment of the present application;

[0188] Fig.17 A schematic diagram of a flow chart of a method for adjusting a first loading strategy provided in an embodiment of the present application;

[0189] Fig.18 A schematic diagram of the structure of a model loading strategy determination device provided in an embodiment of the present application;

[0190] Fig.19 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0191] In order to make the purpose, technical solution and beneficial effects of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0192] refer to Figure 1 , which is a system architecture diagram for determining a model loading strategy applicable to an embodiment of the present application. The system architecture diagram for determining a model loading strategy at least includes a terminal device 101 and a model loading strategy determination system 102.

[0193] The terminal device 101 is installed with a target application for determining the model loading strategy, which may be a pre-installed client, a web application, or a small program embedded in other applications, etc. The terminal device 101 may be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited thereto.

[0194] The model loading strategy determination system 102 is a background server of the target application, which is used to provide services for the target application. The model loading strategy determination system 102 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0195] The terminal device 101 and the model loading strategy determination system 102 may be directly or indirectly connected via wired or wireless communication, which is not limited in the present application.

[0196] The terminal device 101 responds to the user's operation of determining the model loading strategy and sends a model loading strategy determination instruction to the model loading strategy determination system 102 . The model loading strategy determination system 102 receives the model loading strategy determination instruction, obtains the outbound call files corresponding to n1 outbound call scenarios respectively, wherein the outbound call files include the outbound call numbers to be dialed; n1 is an integer greater than 0; then, the outbound call models corresponding to the n1 outbound call scenarios are determined, and the model estimated throughputs corresponding to the n1 outbound call models are determined according to the outbound call numbers in the n1 outbound call files; thereafter, the first loading strategy is determined according to the model estimated throughputs corresponding to the n1 outbound call models and the process expected throughput that a process can provide, wherein the first loading strategy includes the number of processes m1 to be generated, n2 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each process; the process expected throughput is used to indicate that the load of a process is in the optimal state; the n1 outbound call models include n2 target outbound call models; n2 is less than or equal to n1, and m1 and n2 are both integers greater than 0.

[0197] based on Figure 1 The system architecture diagram, the embodiment of the present application provides a process of a method for determining a model loading strategy, such as Figure 2 As shown, the process of this method is Figure 1 The model loading strategy determination system 102 shown is executed, including the following steps:

[0198] S201, obtaining outbound call files corresponding to n1 outbound call scenarios respectively, where n1 is an integer greater than 0.

[0199] In an embodiment of the present application, outbound call scenarios may include marketing scenarios, collection scenarios, etc., wherein the marketing scenarios tend to conduct product marketing and product communication with corporate customers, and the collection scenarios tend to urge corporate customers to repay.

[0200] In a possible implementation, one outbound call scenario may correspond to one outbound call file or multiple outbound call files, which is not limited here. The outbound call file corresponding to any outbound call scenario includes information such as the customer name and outbound call number to be dialed in the scenario.

[0201] It should be understood that the outbound call file can be in XML format, TXT format, or other formats, which are not limited here.

[0202] S202, determining outbound call models corresponding to n1 outbound call scenarios respectively.

[0203] In the above S202, different outbound call scenarios correspond to different outbound call models. For example, a marketing scenario corresponds to a marketing model, wherein the marketing model is an outbound call model trained with marketing corpus; a collection scenario corresponds to a collection model, wherein the collection model is an outbound call model trained with collection corpus.

[0204] S203: Determine the model estimated throughputs corresponding to the n1 outbound call models respectively according to the outbound call numbers in the n1 outbound call files.

[0205] In the above S203, for the i-th outbound call model among the n1 outbound call models, where i is an integer between [1, n1], the following can be used: Figure 3 The following steps shown determine the model estimated throughput corresponding to the i-th outbound call model.

[0206] S301, determining a current outbound call window and a historical outbound call window corresponding to the current outbound call window.

[0207] In the above 301, the current outbound call window is a time window in the current date, wherein the current date may be the date when the above model loading strategy determination method is executed. For example, the current outbound call window is 9:00-12:00 on December 16, 2024.

[0208] The historical outbound call window is a time window in any date before the current date. Generally speaking, the historical outbound call window is a time window in a date before and immediately before the current date. For the convenience of explanation below, the historical outbound call windows mentioned below are all time windows in a date before and immediately before the current date. For example, the historical outbound call window is 9:00-12:00 on December 15, 2024.

[0209] It should be understood that in order to improve the accuracy of the determined model loading strategy, the present application sets the length of the current outbound call window to be the same as that of the historical outbound call window, and the time period of the current outbound call window and the historical outbound call window is also the same.

[0210] In the embodiment of the present application, one current outbound call window may be determined, or multiple current outbound call windows may be determined, which is not limited here.

[0211] S302: Determine the outbound call number in the current outbound call window from the ith outbound call file.

[0212] In the above S302, the i-th outbound call file and the i-th outbound call model correspond to the same outbound call scenario.

[0213] In a possible implementation, the outbound calling numbers in the ith outbound calling file are allocated to each time period, and the outbound calling numbers located in the current outbound calling window are selected from the outbound calling numbers allocated to each time period.

[0214] It should be understood that the outbound call numbers allocated to each time period can be selected through different current outbound call windows, and this application takes one of the current outbound call windows as an example for explanation.

[0215] S303: Determine a historical outbound call number within the historical outbound call window from the historical outbound call file corresponding to the i-th outbound call file.

[0216] In the above S303, the ith outbound call file and the historical outbound call file corresponding to the ith outbound call file correspond to the same outbound call scenario. The historical outbound call file corresponding to the ith outbound call file includes information such as the historical customer name to be dialed and the historical outbound call number.

[0217] It should be understood that the method for determining the historical outbound call number in the historical outbound call window is the same as the method for determining the outbound call number in the current outbound call window, and is not limited here.

[0218] S304: Determine a model estimated throughput corresponding to the ith outbound call model according to the outbound call number in the current outbound call window and the historical outbound call number in the historical outbound call window.

[0219] In the above S304, the Figure 4 The following steps shown determine the model estimated throughput corresponding to the i-th outbound call model.

[0220] S401, determining the number of outbound call numbers corresponding to the current outbound call window according to the outbound call numbers in the current outbound call window.

[0221] In the above S401, the number of outbound call numbers in the current outbound call window is directly used as the number of outbound call numbers corresponding to the current outbound call window.

[0222] S402: Determine a historical answering rate and a historical average number of interactions corresponding to the historical outbound calling window according to the historical outbound calling number in the historical outbound calling window.

[0223] In the above S402, the historical answering rate corresponding to the historical outbound calling window is equal to the ratio of the historical answering number corresponding to the historical outbound calling window to the historical outbound calling number number corresponding to the historical outbound calling window, wherein the historical outbound calling number number corresponding to the historical outbound calling window is the number of historical outbound calling numbers located in the historical outbound calling window, and the historical answering number corresponding to the historical outbound calling window is the number of answered calls among the historical outbound calling numbers located in the historical outbound calling window.

[0224] The historical average number of interactions corresponding to the historical outbound calling window is equal to the ratio of the historical total number of interactions corresponding to the historical outbound calling window to the number of historical outbound calling numbers corresponding to the historical outbound calling window, wherein the historical total number of interactions corresponding to the historical outbound calling window is the sum of the interaction numbers of the historical outbound calling numbers located in the historical outbound calling window, and the interaction number of the historical outbound calling numbers located in the historical outbound calling window is the number of question-and-answer interactions between the corporate customers of the historical outbound calling numbers located in the historical outbound calling window and the outbound calling model.

[0225] It should be understood that a question-and-answer interaction is understood as follows: after the outbound model dials a historical outbound number and the corporate customer of the historical outbound number is answered, the outbound model asks the corporate customer a question, and the corporate customer's reply to the outbound model's question can be regarded as a question-and-answer interaction; or, the corporate customer asks the outbound model a question, and the outbound model replies to the corporate customer's question can be regarded as a question-and-answer interaction.

[0226] S403, determining a model estimated throughput corresponding to the ith outbound call model according to the number of outbound call numbers corresponding to the current outbound call window, the historical answering rate corresponding to the historical outbound call window, and the historical average number of interactions.

[0227] In the above S403, the model estimated throughput T(i) corresponding to the i-th outbound call model can be determined by formula (1).

[0228]

[0229] Wherein, T(i) represents the model estimated throughput corresponding to the i-th outbound call model, c represents the number of outbound call numbers corresponding to the current outbound call window, r represents the historical answering rate corresponding to the historical outbound call window, a represents the historical average number of interactions corresponding to the historical outbound call window, and len_window represents the length of the current outbound call window.

[0230] It should be understood that the model estimated throughput corresponding to the i-th outbound call model may also be the transaction processing throughput corresponding to the i-th outbound call model, such as (Transactions Per Second, TPS for short), or other throughputs, which are not limited here.

[0231] Figure 4 The steps shown provide a method for specifically determining the model estimated throughput corresponding to the i-th outbound call model.

[0232] S204: Determine a first loading strategy according to the model estimated throughputs respectively corresponding to the n1 outbound call models and the process expected throughput that can be provided by one process.

[0233] In an embodiment of the present application, the first loading strategy includes the number of processes m1 to be generated, the n2 target outbound call models to be loaded in each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each of the m1 processes; the n1 outbound call models include n2 target outbound call models; n2 is less than or equal to n1, and m1 and n2 are both integers greater than 0.

[0234] In the embodiment of the present application, in order to ensure that the outbound call model can stably provide outbound call services, the n2 target outbound call models to be loaded by different processes may overlap, that is, any target outbound call model can be loaded by at least two processes. When a target outbound call model is loaded by at least two processes, each process can allocate a certain process throughput to the target outbound call model, and the sum of the process throughput allocated by each process to the target outbound call model is the model estimated throughput corresponding to the target outbound call model.

[0235] It should be understood that the expected process throughput that a process can provide may be the transaction processing throughput that a process can provide, or may be other throughputs, which is not limited here.

[0236] In the embodiment of the present application, the expected process throughput that a process can provide can be obtained by measuring the process, can be determined with reference to historical experience, or can be set manually, which is not limited here.

[0237] For the convenience of the following description, the expected throughput that a process can provide can be marked as exp_worker_tps.

[0238] In an embodiment of the present application, the model maximum throughput max_model_tps, the model minimum throughput min_model_tps, the process maximum throughput max_worker_tps, and the process minimum throughput min_worker_tps can be determined according to the expected process throughput exp_worker_tps, wherein the model maximum throughput max_model_tps is the upper limit of the throughput allocated by a process to an outbound call model, the model minimum throughput min_model_tps is the lower limit of the throughput allocated by a process to an outbound call model, the process maximum throughput max_worker_tps is the upper limit of the throughput that a process can provide, and the process minimum throughput min_worker_tps is the lower limit of the throughput that a process can provide.

[0239] It should be understood that the maximum process throughput max_worker_tps and the minimum process throughput min_worker_tps may constitute a process throughput range of a process, namely [max_worker_tps, min_worker_tps].

[0240] In the embodiment of the present application, the model maximum throughput max_model_tps can be determined according to the process expected throughput exp_worker_tps. Specifically, the model maximum throughput max_model_tps can be determined by formula (2).

[0241]

[0242] Among them, max_model_tps indicates the maximum model throughput, exp_worker_tps indicates the expected throughput that a process can provide, and num1 indicates the number of outbound call models that a process needs to load at least. Generally speaking, the value of num1 is greater than 0. For example, a value of num1 of 2 indicates that a process needs to load at least 2 outbound call models. When a process needs to load at least 2 outbound call models, the throughput of each outbound call model in the process will not exceed half of the expected throughput of the process.

[0243] In the embodiment of the present application, the model minimum throughput min_model_tps can be determined according to the process expected throughput exp_worker_tps and the model estimated throughputs corresponding to the n1 outbound call models. Specifically, the model minimum throughput min_model_tps can be determined by formula (3).

[0244]

[0245] Among them, min_model_tps represents the minimum throughput of the model, T(i) represents the model estimated throughput corresponding to the i-th outbound call model, min(T(i)) represents the minimum value of the model estimated throughput corresponding to the n1 outbound call models, max_model_tps represents the maximum throughput of the model, and num2 represents the minimum number of processes required for an outbound call model to be loaded. Generally speaking, in order to ensure that the outbound call model can provide outbound call services stably, the value of num2 is greater than 1, that is, an outbound call model is loaded by at least multiple processes, so that even if one of the processes fails, the outbound call model in other processes is still available. For example, the value of num2 is 2, indicating that an outbound call model is loaded by at least 2 processes. For the convenience of discussion, the value of num2 can be the same as the value of num1. For the convenience of explanation, the values ​​of num1 and num2 are the same in the following text, and will not be repeated in the following text. In an embodiment of the present application, the maximum throughput of the process max_worker_tps can be determined based on the expected throughput of the process exp_worker_tps and the minimum throughput of the model min_model_tps. Specifically, the maximum process throughput max_worker_tps can be determined by formula (4).

[0246]

[0247] Among them, max_worker_tps indicates the maximum throughput of the process, exp_worker_tps indicates the expected throughput that a process can provide, min_model_tps indicates the minimum throughput of the model, and num2 indicates the minimum number of processes required to load an outbound call model.

[0248] In the embodiment of the present application, the minimum process throughput min_worker_tps can be determined according to the expected process throughput exp_worker_tps and the model minimum throughput min_model_tps. Specifically, the minimum process throughput min_worker_tps can be determined by formula (5).

[0249]

[0250] Among them, min_worker_tps indicates the minimum process throughput, exp_worker_tps indicates the expected process throughput that a process can provide, min_model_tps indicates the minimum model throughput, and num2 indicates the minimum number of processes required to load an outbound call model.

[0251] It should be understood that the model maximum throughput max_model_tps, model minimum throughput min_model_tps, process maximum throughput max_worker_tps, and process minimum throughput min_worker_tps determined by the above formula are all used in the following text and will not be repeated herein.

[0252] For example, it is assumed that there are 10 outbound call scenarios and outbound call models corresponding to the 10 outbound call scenarios. Through S203, the model estimated throughputs corresponding to the 10 outbound call models can be determined as shown in Table 1.

[0253] Table 1.

[0254]

[0255]

[0256] Set the expected throughput exp_worker_tps that a process can provide to 400, the number of outbound call models that a process needs to load at least num1 to 2, and the minimum number of processes num2 required to load an outbound call model to 2, and then determine the maximum model throughput max_model_tps = 400 / 2 = 200, and the minimum model throughput The maximum process throughput max_worker_tps = 400 + (56 / 2) = 428, and the minimum process throughput min_worker_tps = 400 - (56 / 2) = 372. The process throughput range of a process is [372, 428].

[0257] In the above method, first, the outbound call files corresponding to n1 outbound call scenarios are obtained, wherein the outbound call files include the outbound call numbers to be dialed; then, the outbound call models corresponding to the n1 outbound call scenarios are determined, and according to the outbound call numbers in the n1 outbound call files, the model estimated throughputs corresponding to the n1 outbound call models are determined; then, according to the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that can be provided by a process, the first loading strategy is determined, wherein the first loading strategy includes the number of processes m1 to be generated, the n2 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each of the m1 processes. The above model loading process does not directly load all outbound call models into one process, nor does it directly load one outbound call model into one process. Instead, it uses the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that one process can provide to accurately determine the number of processes m1 required to load the above n1 outbound call models, the target outbound call model to be loaded by each of the m1 processes, and the process throughput allocated to each target outbound call model by each of the m1 processes. The model loading strategy determined by the above method fully considers the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that one process can provide. The determined model loading strategy is more reasonable and accurate, which can effectively reduce the waste of process load resources.

[0258] Optionally, in the above S204, the following Figure 5 The following steps shown determine a first loading strategy.

[0259] S501, determining the model residual throughputs respectively corresponding to the n1 outbound call models according to the model estimated throughputs respectively corresponding to the n1 outbound call models.

[0260] In the above S501, the model estimated throughputs corresponding to the n1 outbound call models are the throughputs required to be allocated corresponding to the n1 outbound call models, and the model remaining throughputs corresponding to the n1 outbound call models are the throughputs still required to be allocated corresponding to the n1 outbound call models.

[0261] Since the process throughput has not been allocated to each outbound call model currently, the model estimated throughputs corresponding to the n1 outbound call models respectively can be directly used as the model remaining throughputs corresponding to the n1 outbound call models respectively.

[0262] S502, repeatedly execute the following process until the model remaining throughputs corresponding to the n1 outbound call models are 0: determine a process; select n2 target outbound call models from the n1 outbound call models, and determine the process throughputs respectively allocated by the process to the n2 target outbound call models based on the expected process throughputs; use the process throughputs respectively allocated by the process to the n2 target outbound call models to update the model remaining throughputs respectively corresponding to the n1 outbound call models.

[0263] Optionally, in the above 502, n2 target outbound call models are selected from the n1 outbound call models, and the process throughputs allocated by one process to the n2 target outbound call models are determined according to the expected process throughputs. The process throughputs allocated by one process to the n2 target outbound call models are used to update the model residual throughputs corresponding to the n1 outbound call models. Figure 6 The following steps are shown to achieve this.

[0264] S601, determining the allocated throughput corresponding to the process, and initializing the allocated throughput corresponding to the process.

[0265] In the above S601, since the process currently allocates process throughput to any outbound call model, the allocated throughput corresponding to the process may be initialized to 0.

[0266] Optionally, the throughput to be allocated corresponding to the process may also be determined according to the minimum throughput of the process and the allocated throughput corresponding to the process. Specifically, the difference between the minimum throughput of the process min_worker_tps and the allocated throughput corresponding to the process is the throughput to be allocated corresponding to the process.

[0267] S602, repeatedly perform the following steps until the allocated throughput corresponding to the process is within the process throughput range: sort the model remaining throughputs corresponding to the n1 outbound call models respectively according to the sorting rule to obtain a first sequence; determine the target outbound call model according to the model remaining throughput corresponding to the outbound call model located at the first position in the first sequence; determine the process throughput allocated by the process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model; use the process throughput allocated by the process to the target outbound call model to update the allocated throughput corresponding to the process, and update the model remaining throughputs corresponding to the n1 outbound call models respectively.

[0268] In the above S602, the first sequence can be obtained by any one of implementation modes A1, A2, A3, and A4.

[0269] Implementation method A1, sorting the model remaining throughputs corresponding to the n1 outbound call models respectively in ascending order to obtain a first sequence.

[0270] In the above-mentioned implementation A1, the first sequence includes model residual throughputs corresponding to n1 outbound call models respectively arranged in ascending order.

[0271] Implementation A1 provides a method for conveniently obtaining a first sequence.

[0272] Implementation method A2, sorting the model remaining throughputs corresponding to the n1 outbound call models respectively in descending order to obtain a first sequence.

[0273] In the above-mentioned implementation A2, the first sequence includes the model remaining throughputs corresponding to n1 outbound call models respectively arranged in descending order.

[0274] Implementation A2 provides a method for conveniently obtaining a first sequence.

[0275] Embodiment A3, comprising Fig. 7A The following steps are shown.

[0276] S701A, determining the number S of determined outbound call models, wherein the determined outbound call model is a target outbound call model corresponding to the allocated throughput corresponding to the process.

[0277] In the above S701A, the process is the sum of the process throughputs respectively allocated to the multiple determined outbound call models, that is, the allocated throughput corresponding to the process. The number S of the determined outbound call models can be determined according to the allocated throughput corresponding to the process.

[0278] S702A, when it has been determined that the number S of outbound call models is an even number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in descending order to obtain a first sequence.

[0279] In S702A, the first sequence includes model residual throughputs corresponding to n1 outbound call models arranged in descending order.

[0280] S703A, when it has been determined that the number S of outbound call models is an odd number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in ascending order to obtain a first sequence.

[0281] In S703A, the first sequence includes model residual throughputs corresponding to n1 outbound call models arranged in ascending order.

[0282] Implementation A3 provides a method for obtaining a first sequence. The multiple first sequences obtained by Implementation A3 are arranged alternately in ascending and descending order, which is conducive to subsequently allocating an outbound call model to different processes.

[0283] Embodiment A4, comprising Figure 7B The following steps are shown.

[0284] S701B, determining the number S of determined outbound call models, wherein the determined outbound call model is a target outbound call model corresponding to the allocated throughput corresponding to the process.

[0285] S702B, when it has been determined that the number S of outbound call models is an odd number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in descending order to obtain a first sequence.

[0286] In S702B, the first sequence includes model remaining throughputs corresponding to n1 outbound call models arranged in descending order.

[0287] S703B, when it has been determined that the number S of outbound call models is an even number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in ascending order to obtain a first sequence.

[0288] In S703B, the first sequence includes model residual throughputs corresponding to n1 outbound call models arranged in ascending order.

[0289] Implementation A4 provides a method for obtaining a first sequence. The multiple first sequences obtained by Implementation A4 are arranged alternately in ascending and descending order, which is conducive to subsequently allocating an outbound call model to different processes.

[0290] To facilitate explanation below, implementation mode A3 is used below to obtain the first sequence.

[0291] In the above S602, the target outbound call model is determined according to the model remaining throughput corresponding to the outbound call model at the first position in the first sequence, which can be implemented by any one of implementation modes B1, B2, and B3.

[0292] In implementation mode B1, when the first sequence is obtained through implementation mode A1, the model remaining throughput corresponding to the outbound call model at the first position in the first sequence is the minimum value of the model remaining throughputs corresponding to the n1 outbound call models respectively, and the target outbound call model is the outbound call model corresponding to the minimum value.

[0293] In implementation mode B2, when the first sequence is obtained through implementation mode A2, the model remaining throughput corresponding to the outbound call model at the first position in the first sequence is the maximum value of the model remaining throughputs corresponding to the n1 outbound call models, and the target outbound call model is the outbound call model corresponding to the maximum value.

[0294] Implementation method B3, when the first sequence is obtained through implementation method A3 or A4, if the first sequence includes the model residual throughputs corresponding to n1 outbound call models arranged in descending order, then the model residual throughput corresponding to the outbound call model located at the first position in the first sequence is the maximum value among the model residual throughputs corresponding to the n1 outbound call models, and the target outbound call model is the outbound call model corresponding to the maximum value. If the first sequence includes the model residual throughputs corresponding to n1 outbound call models arranged in ascending order, then the model residual throughput corresponding to the outbound call model located at the first position in the first sequence is the minimum value among the model residual throughputs corresponding to the n1 outbound call models, and the target outbound call model is the outbound call model corresponding to the minimum value.

[0295] To facilitate the explanation below, implementation method B3 is used below to determine the target outbound call model.

[0296] In the above S602, the process throughput allocated to the target outbound call model can be determined according to the model remaining throughput, the model maximum throughput max_model_tps, and the model minimum throughput min_model_tps corresponding to the target outbound call model. Figure 8 The following steps are shown to achieve this.

[0297] S801, when the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than or equal to a first preset value, determine the process throughput allocated to the target outbound call model by the process according to the model remaining throughput or the model minimum throughput corresponding to the target outbound call model.

[0298] In the embodiment of the present application, the first preset value is the minimum number of processes num2 required to load an outbound call model, which will not be described in detail below.

[0299] In the above S801, the following two branches may be included, namely, branch 1 and branch 2.

[0300] Branch 1, when the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than a first preset value, the model remaining throughput corresponding to the target outbound call model is used as the process throughput allocated by the process to the target outbound call model.

[0301] In the above branch 1, when the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than the first preset value, it means that the model remaining throughput corresponding to the target outbound call model can no longer be split. If the model remaining throughput corresponding to the target outbound call model is still split, the model remaining throughput corresponding to the split target outbound call model is less than the model minimum throughput, which is not conducive to model loading.

[0302] Branch 2: When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is equal to the first preset value, the model minimum throughput is used as the process throughput allocated by the process to the target outbound call model.

[0303] In the above branch 2, when the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is equal to the first preset value, it means that the model remaining throughput corresponding to the target outbound call model can be split into num2. Therefore, the model minimum throughput can be used as the process throughput allocated by the process to the target outbound call model.

[0304] S802, when the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, determine the process throughput allocated to the target outbound call model by the process according to the model remaining throughput and the model maximum throughput corresponding to the target outbound call model.

[0305] In the above S802, the following two branches may be included, namely branch 3 and branch 4.

[0306] Branch 3, when the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than the first preset value, if the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, the process throughput allocated to the target outbound call model by the process is determined according to the model remaining throughput corresponding to the target outbound call model, the throughput to be allocated corresponding to the process, and the model minimum throughput.

[0307] In the above branch 3, the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, which means that all the model remaining throughput corresponding to the target outbound call model can be allocated, that is, all the model remaining throughput corresponding to the target outbound call model is allocated to the process and does not exceed the model maximum throughput.

[0308] In the above branch 3, it can also be divided into two branches, namely branch 3.1 and branch 3.2.

[0309] Branch 3.1: If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the process and the process throughput control value, then the process throughput allocated to the target outbound call model by the process is determined according to the model remaining throughput and the model minimum throughput corresponding to the target outbound call model.

[0310] In the embodiment of the present application, the process throughput control value is determined according to the process throughput range. Specifically, the difference between the maximum process throughput max_worker_tps and the minimum process throughput min_worker_tps is the process throughput control value.

[0311] In an embodiment of the present application, the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the process and the process throughput control value, indicating that the model remaining throughput corresponding to the target outbound call model is fully allocated to the process and will not exceed the process maximum throughput of the process.

[0312] The above branch 3.1 can be split into two branches, namely branch 3.1.1 and branch 3.1.2.

[0313] Branch 3.1.1: If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is not equal to the model estimated throughput corresponding to the target outbound call model, then the model remaining throughput corresponding to the target outbound call model is used as the process throughput allocated by the process to the target outbound call model.

[0314] In an embodiment of the present application, if the model remaining throughput corresponding to the target outbound call model is not equal to the model estimated throughput corresponding to the target outbound call model, it means that the model remaining throughput corresponding to the target outbound call model has participated in process allocation; otherwise, it means that the model remaining throughput corresponding to the target outbound call model has not participated in process allocation.

[0315] Branch 3.1.2: If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is equal to the model estimated throughput corresponding to the target outbound call model, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated to the target outbound call model by the process.

[0316] Branch 3.2: If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the process and the process throughput control value, then the process throughput allocated to the target outbound call model by the process is determined according to the throughput to be allocated corresponding to the process and the model minimum throughput.

[0317] In an embodiment of the present application, the model remaining throughput corresponding to the target outbound call model is greater than the sum of the to-be-allocated throughput corresponding to the process and the process throughput control value, indicating that allocating all the model remaining throughput corresponding to the target outbound call model to the process will exceed the maximum process throughput of the process.

[0318] In the above branch 3.2, the process throughput allocated to the target outbound call model by the process is determined according to the to-be-allocated throughput corresponding to the process and the minimum throughput of the model. This can be specifically achieved in the following way: the maximum value of the to-be-allocated throughput corresponding to the process and the minimum throughput of the model is selected as the process throughput allocated to the target outbound call model by the process.

[0319] Branch 4, when the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than the first preset value, if the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, then the process throughput allocated to the target outbound call model by the process is determined according to the model remaining throughput, the model maximum throughput and the model minimum throughput corresponding to the target outbound call model.

[0320] The above branch 4 can also be divided into two branches, namely branch 4.1 and branch 4.2.

[0321] Branch 4.1: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, then the process throughput allocated to the target outbound call model by the process is determined according to the model remaining throughput corresponding to the target outbound call model, the model minimum throughput, and the throughput to be allocated corresponding to the process.

[0322] In the embodiment of the present application, the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is smaller than the model minimum throughput, indicating that allocation cannot be performed according to the model maximum throughput. If allocation is performed according to the model maximum throughput, the model remaining throughput corresponding to the target outbound call model after allocation will be smaller than the model minimum throughput, i.e., allocation cannot be performed again.

[0323] The above branch 4.1 can also be split into two branches, namely branch 4.1.1 and branch 4.1.2.

[0324] Branch 4.1.1: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is less than the throughput to be allocated corresponding to the process, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated to the target outbound call model by the process.

[0325] In the embodiment of the present application, the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is smaller than the throughput to be allocated corresponding to the process, indicating that allocation can be performed according to the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput. The model remaining throughput corresponding to the target outbound call model after allocation is the model minimum throughput, and can be allocated again.

[0326] Branch 4.1.2: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is greater than or equal to the throughput to be allocated corresponding to the process, then the maximum value between the throughput to be allocated corresponding to the process and the model minimum throughput is used as the process throughput allocated to the target outbound call model by the process.

[0327] Branch 4.2: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, then the process throughput allocated to the target outbound call model by the process is determined according to the model minimum throughput, the model maximum throughput, and the throughput to be allocated corresponding to the process.

[0328] In the embodiment of the present application, the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, indicating that if allocation is performed according to the model maximum throughput, the model remaining throughput corresponding to the target outbound call model after allocation will be greater than or equal to the model minimum throughput, and allocation can be performed again later.

[0329] The above branch 4.2 can also be split into two branches, namely branch 4.2.1 and branch 4.2.2.

[0330] Branch 4.2.1: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput and the model maximum throughput corresponding to the target outbound call model is greater than or equal to the model minimum throughput, and the model maximum throughput is less than or equal to the sum of the to-be-allocated throughput corresponding to the process and the process throughput control value, then the model maximum throughput is used as the process throughput allocated by the process for the target outbound call model.

[0331] Branch 4.2.2: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is greater than the sum of the to-be-allocated throughput corresponding to a process and the process throughput control value, then the sum of the to-be-allocated throughput corresponding to a process and the process throughput control value is used as the process throughput allocated to the target outbound call model as a process.

[0332] It should be understood that the priorities of the above-mentioned branch 1, branch 2, branch 3, and branch 4 are in the order of branch 1, branch 2, branch 3, and branch 4 from high to low.

[0333] Among them, branch 3 includes branch 3.1 and branch 3.2, and the priority of branch 3.1 is higher than that of branch 3.2. Branch 3.1 includes branch 3.1.1 and branch 3.1.2, and the priority of branch 3.1.1 is higher than that of branch 3.1.2.

[0334] Among them, branch 4 includes branch 4.1 and branch 4.2, and the priority of branch 4.1 is higher than that of branch 4.2. Branch 4.1 includes branch 4.1.1 and branch 4.1.2, and the priority of branch 4.1.1 is higher than that of branch 4.1.2. Branch 4.2 includes branch 4.2.1 and branch 4.2.2, and the priority of branch 4.2.1 is higher than that of branch 4.2.2.

[0335] In the above S602, the process throughput allocated by the process to the target outbound call model is used to update the allocated throughput corresponding to the process, and the model remaining throughputs corresponding to the n1 outbound call models are updated. This can be done by: Fig. 9 The following steps are implemented.

[0336] S901, the allocated throughput corresponding to the process is added to the process throughput allocated by the process for the target outbound call model, that is, the allocated throughput corresponding to the process after update.

[0337] S902: Subtract the process throughput allocated by the process to the target outbound call model from the model remaining throughput corresponding to the target outbound call model to obtain the model remaining throughput corresponding to the updated target outbound call model.

[0338] For example, the model estimated throughput corresponding to 10 outbound call models is shown in Table 1. The expected process throughput exp_worker_tps that a process can provide is 400. The minimum number of outbound call models that a process needs to load is num1, which is 2. The minimum number of processes required to load an outbound call model is num2, which is 2.

[0339] Based on the above information, the model maximum throughput max_model_tps, the model minimum throughput min_model_tps, the process maximum throughput max_worker_tps, the process minimum throughput min_worker_tps, the process throughput range of a process, and the process throughput control value of a process can be determined as shown in Table 2.

[0340] Table 2.

[0341] Model maximum throughput max_model_tps 200 Model minimum throughput min_model_tps 56 Maximum throughput of the process max_worker_tps 428 Process minimum throughput min_worker_tps 372 The process throughput range of a process [372,428] The process throughput control value of a process 56

[0342] According to the model estimated throughputs corresponding to the 10 outbound call models in Table 1, the model residual throughputs corresponding to the 10 outbound call models in Table 3 can be determined. Since the process throughput has not yet been allocated to each outbound call model, the model residual throughput corresponding to the outbound call model is the model estimated throughput corresponding to the outbound call model.

[0343] Table 3.

[0344] Model No. 1 2 3 4 5 6 7 8 9 10 Model Estimated Throughput 970 911 666 654 567 453 368 287 261 112 Model Remaining Throughput 970 911 666 654 567 453 368 287 261 112

[0345] A process is determined, and the identifier of the process is process 1. The allocated throughput corresponding to process 1 is initialized to 0, and the throughput to be allocated corresponding to process 1 is 372 (ie, 372-0=372).

[0346] Since the allocated throughput corresponding to process 1 is 0, which is not within the process throughput range [372,428], perform the following steps:

[0347] It is determined that the number S of the determined outbound call models corresponding to process 1 is 0. Since 0 is an even number, the model residual throughputs corresponding to the 10 outbound call models are arranged in descending order to obtain a first sequence, wherein the first sequence can refer to Table 3. The maximum value of the model residual throughput corresponding to the outbound call model is selected from the first sequence, that is, 970, and the outbound call model corresponding to 970 is the outbound call model 1.

[0348] According to the calculation, it is determined that branch 4.2.1 in the above branch is met, that is, the following conditions are met:

[0349] (1) The quotient of the model's remaining throughput (i.e., 970) and the model's minimum throughput (i.e., 56) is greater than num2 (i.e., 2)

[0350] (2) The model's remaining throughput (i.e., 970) is greater than the model's maximum throughput (i.e., 200)

[0351] (3) The difference between the model's remaining throughput (i.e., 970) and the model's maximum throughput (i.e., 200) is greater than the model's minimum throughput (i.e., 56)

[0352] (4) The maximum throughput of the model (i.e., 200) is less than the sum of the throughput to be allocated corresponding to process 1 (i.e., 372) and the process throughput control value (i.e., 56)

[0353] Therefore, the model maximum throughput (i.e., 200) is used as the process throughput allocated by process 1 for outbound call model 1. The allocated throughput corresponding to updated process 1 is 200 (0+200=200). The throughput to be allocated corresponding to updated process 1 is 172 (i.e., 372-200=172). The model remaining throughput corresponding to updated outbound call model 1 is 770 (970-200=770).

[0354] At this time, the model remaining throughputs corresponding to the 10 outbound call models can be referred to Table 4.

[0355] Table 4.

[0356] Model No. 1 2 3 4 5 6 7 8 9 10 Model Estimated Throughput 970 911 666 654 567 453 368 287 261 112 Model Remaining Throughput 770 911 666 654 567 453 368 287 261 112

[0357] Since the allocated throughput of process 1 is 200, which is not within the process throughput range [372,428], perform the following steps:

[0358] It is determined that the number S of the determined outbound call models corresponding to process 1 is 1. Since 1 is an odd number, the model residual throughputs corresponding to the 10 outbound call models are arranged in ascending order to obtain a first sequence, wherein the first sequence can refer to Table 5. The minimum value of the model residual throughput corresponding to the outbound call model is selected from the first sequence, that is, 112, and the outbound call model corresponding to 112 is the outbound call model 10.

[0359] Table 5.

[0360] Model No. 10 9 8 7 6 5 4 3 1 2 Model Estimated Throughput 112 261 287 368 453 567 654 666 970 911 Model Remaining Throughput 112 261 287 368 453 567 654 666 770 911

[0361] According to the calculation, it is determined that branch 2 in the above branch is satisfied, that is, the following conditions are satisfied:

[0362] (1) The quotient of the model's remaining throughput (i.e. 112) and the model's minimum throughput (i.e. 56) is equal to num2 (i.e. 2)

[0363] Therefore, the model minimum throughput (i.e., 56) is used as the process throughput allocated by process 1 for outbound call model 10. The allocated throughput corresponding to updated process 1 is 256 (200+56=256). The throughput to be allocated corresponding to updated process 1 is 116 (i.e., 372-256=116). The model remaining throughput corresponding to updated outbound call model 10 is 56 (112-56=56).

[0364] At this time, the model remaining throughputs corresponding to the 10 outbound call models can be referred to Table 6.

[0365] Table 6.

[0366] Model No. 1 2 3 4 5 6 7 8 9 10 Model Estimated Throughput 970 911 666 654 567 453 368 287 261 112 Model Remaining Throughput 770 911 666 654 567 453 368 287 261 56

[0367] Since the allocated throughput of process 1 is 256, which is not within the process throughput range [372,428], perform the following steps:

[0368] It is determined that the number S of the determined outbound call models corresponding to process 1 is 2. Since 2 is an even number, the model residual throughputs corresponding to the 10 outbound call models are arranged in descending order to obtain a first sequence, wherein the first sequence can refer to Table 7. The maximum value of the model residual throughput corresponding to the outbound call model is selected from the first sequence, that is, 911, and the outbound call model corresponding to 911 is the outbound call model 2.

[0369] Table 7.

[0370] Model No. 2 1 3 4 5 6 7 8 9 10 Model Estimated Throughput 911 970 666 654 567 453 368 287 261 112 Model Remaining Throughput 911 770 666 654 567 453 368 287 261 56

[0371] According to the calculation, it is determined that branch 4.2.2 in the above branch is met, that is, the following conditions are met:

[0372] (1) The quotient of the model's remaining throughput (i.e., 911) and the model's minimum throughput (i.e., 56) is greater than num2 (i.e., 2)

[0373] (2) The model's remaining throughput (i.e., 911) is greater than the model's maximum throughput (i.e., 200)

[0374] (3) The difference between the model's remaining throughput (i.e., 911) and the model's maximum throughput (i.e., 200) is greater than the model's minimum throughput (i.e., 56)

[0375] (4) The maximum throughput of the model (i.e., 200) is greater than the sum of the throughput to be allocated corresponding to process 1 (i.e., 116) and the process throughput control value (i.e., 56)

[0376] Therefore, the sum of the throughput to be allocated corresponding to process 1 (i.e., 116) and the process throughput control value (i.e., 56) is taken as the process throughput (i.e., 172) allocated by process 1 for outbound call model 2. The allocated throughput corresponding to updated process 1 is 428 (200+56+172=428). The model remaining throughput corresponding to updated outbound call model 2 is 739 (911-172=739).

[0377] At this time, the model remaining throughputs corresponding to the 10 outbound call models can be referred to Table 8.

[0378] Table 8.

[0379] Model No. 2 1 3 4 5 6 7 8 9 10 Model Estimated Throughput 911 970 666 654 567 453 368 287 261 112 Model Remaining Throughput 739 770 666 654 567 453 368 287 261 56

[0380] Since the allocated throughput corresponding to process 1 is 428, which is within the process throughput range [372, 428], the iteration stops. It can be determined that the target outbound call models to be loaded by process 1 include outbound call model 1, outbound call model 10 and outbound call model 2. It can also be determined that the process throughputs allocated by process 1 to outbound call model 1, outbound call model 10 and outbound call model 2 are 200, 56 and 172 respectively, as shown in Table 9.

[0381] Table 9.

[0382]

[0383] A process is determined again, and the process of process 1 is repeated until the model remaining throughputs corresponding to the 10 outbound call models are all 0.

[0384] Finally, the first loading strategy can be determined, which includes the number of processes to be generated 13, multiple target outbound call models to be loaded by each process, and the process throughput allocated to each process for the multiple target outbound call models, as shown in Table 10.

[0385] Table 10.

[0386]

[0387]

[0388] In one possible implementation, by Figure 2 After the first loading strategy is determined in the steps shown, the first loading strategy can also be executed. Specifically, according to the first loading strategy, m1 processes are generated, and n2 target outbound call models are loaded for each of the m1 processes, and corresponding process throughputs are respectively allocated to the n2 target outbound call models loaded for each of the m1 processes.

[0389] Optionally, for any outbound call number in the outbound call file, you can Fig.10 The following steps are shown to dial the outbound number.

[0390] S1001, obtaining an outbound call number in an outbound call file, and generating a random number for the outbound call number.

[0391] In a possible implementation, the generated random number may be a decimal between 0 and 1.

[0392] S1002, determine the target outbound call model corresponding to the outbound call number, and determine m3 processes where the target outbound call model corresponding to the outbound call number is located, where m3 is less than or equal to m1.

[0393] In the above S1002, m3 processes can be determined from the first loading strategy, wherein the m3 processes are used to load the target outbound call model corresponding to the outbound call number.

[0394] S1003, selecting a target process from m3 processes according to a random number, and assigning the outbound call number to the target process, so that the target outbound call model corresponding to the outbound call number deployed in the target process dials the outbound call number.

[0395] Fig.10 The steps in provide a method for allocating an outbound call number and dialing the outbound call number.

[0396] In the above S1003, the Fig.11 The following steps shown select the target process from among the m3 processes.

[0397] S1101, determining the process throughputs allocated to the target outbound call model corresponding to the outbound call number by m3 processes.

[0398] In the above S1101, the process throughputs allocated to the target outbound call model corresponding to the outbound call number by m3 processes are determined from the first loading strategy.

[0399] S1102: Determine the value ranges corresponding to the m3 processes respectively according to the process throughputs respectively allocated by the m3 processes to the target outbound call model corresponding to the outbound call number.

[0400] In the above S1102, for the jth process among the m3 processes, where j is an integer between [1, m3], the following is executed: Fig.12 The following steps are shown.

[0401] S1201, determining a process throughput ratio allocated to the j-th process for the target outbound call model corresponding to the outbound call number.

[0402] In the above S1201, the process throughput ratio allocated by the jth process to the target outbound call model corresponding to the outbound call number is equal to the ratio of the process throughput allocated by the jth process to the target outbound call model corresponding to the outbound call number to the sum of the process throughputs allocated by m3 processes to the target outbound call models corresponding to the outbound call number.

[0403] S1202: Determine the value interval corresponding to the j-1th process according to the right endpoint value of the value interval corresponding to the j-1th process and the process throughput ratio allocated by the jth process to the target outbound call model corresponding to the outbound call number.

[0404] In the above S1202, the right endpoint value of the value interval corresponding to the j-1th process is used as the left endpoint value of the value interval corresponding to the jth process; the sum of the left endpoint value of the value interval corresponding to the jth process and the process throughput ratio allocated by the jth process to the target outbound call model corresponding to the outbound call number is used as the right endpoint value of the value interval corresponding to the jth process; based on the left endpoint value and the right endpoint value of the value interval corresponding to the jth process, the value interval corresponding to the jth process can be determined.

[0405] It should be understood that when the value of j is 1, the right endpoint value of the value interval corresponding to the 0th process is 0.

[0406] S1103, selecting a process corresponding to the value interval of the random number as the target process.

[0407] Fig.11 The steps in provide a method for selecting a corresponding target process for an outbound call number.

[0408] In one possible implementation, by Figure 2 After the steps shown determine the first loading strategy, the first loading strategy can also be adjusted, including the following steps: if the allocated throughput corresponding to the m1th process among m1 processes is less than the minimum throughput of the process, the first loading strategy is adjusted to obtain the second loading strategy, wherein the second loading strategy includes the number of processes m2 to be generated, the n3 target outbound call models to be loaded by each of the m2 processes, and the process throughput allocated by each of the m2 processes to the n3 target outbound call models; m2 is less than or equal to m1, and n3 is an integer greater than 0.

[0409] Specifically, the method for adjusting the first loading strategy may include: Fig.13 The following steps are shown.

[0410] S1301, if the allocated throughput corresponding to the m1th process among m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is less than the ratio of the expected throughput of the process to the second preset value, then the allocated throughput corresponding to the m1th process is allocated to other processes to obtain a second loading strategy; the second loading strategy includes the number of processes to be generated (m1)-1, the n3 target outbound call models to be loaded by each of the (m1)-1 processes, and the process throughput allocated to the n3 target outbound call models by each of the (m1)-1 processes.

[0411] It should be understood that the second preset value is a constant value, for example, the second preset value is 2. The other processes are processes other than the m1-th process among the m1 processes.

[0412] In one possible implementation, repeatedly executing Fig.14 The following process is shown until the allocated throughput corresponding to the m1th process is 0:

[0413] S1401, for the kth target outbound call model among the n2 target outbound call models that allocate process throughput to the m1th process, determine m5 processes to be adjusted from other processes, the determined m5 processes to be adjusted are all allocated process throughput for the kth target outbound call model, k is an integer between [1, n2], and m5 is less than or equal to m1.

[0414] S1402, determining throughputs to be adjusted corresponding to m5 processes to be adjusted respectively.

[0415] In the above S1402, for the pth process to be adjusted among the m5 processes to be adjusted, p is an integer between [1, m5], and the throughput to be adjusted corresponding to the pth process to be adjusted can be determined by the following method: select the minimum value from the difference between the maximum throughput of the process and the allocated throughput corresponding to the pth process to be adjusted, and the process throughput allocated by the m1th process to the kth target outbound call model, as the throughput to be adjusted corresponding to the pth process to be adjusted.

[0416] In addition, the allocated throughput corresponding to the pth process to be adjusted may also be updated. Specifically, the allocated throughput corresponding to the pth process to be adjusted is added to the throughput to be adjusted corresponding to the pth process to be adjusted, which is the updated allocated throughput corresponding to the pth process to be adjusted.

[0417] In addition, it can be determined that the target outbound call model to be loaded by the pth process to be adjusted also includes the kth target outbound call model, and the process throughput allocated by the pth process to be adjusted to the kth target outbound call model can also be determined.

[0418] S1403: Update the allocated throughput corresponding to the m1th process according to the throughputs to be adjusted corresponding to the m5 processes to be adjusted.

[0419] In the above S1403, the allocated throughput corresponding to the m1th process is subtracted from the throughputs to be adjusted corresponding to the m5 processes to be adjusted, which is the allocated throughput corresponding to the m1th process after update.

[0420] Fig.14 A first loading strategy adjustment method is provided.

[0421] S1302: If the allocated throughput corresponding to the m1th process among m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is greater than or equal to the ratio of the expected throughput of the process to the second preset value, the allocated throughputs corresponding to other processes are allocated to the m1th process to obtain a second loading strategy; the second loading strategy includes the number of processes m1 to be generated, n3 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated by each of the m1 processes to the n3 target outbound call models.

[0422] In one possible implementation, repeatedly executing Fig.15 The following process is shown until the allocated throughput corresponding to the m1th process is greater than or equal to the minimum throughput of the process:

[0423] S1501, for the kth target outbound call model among the n2 target outbound call models that allocate process throughput to the m1th process, determine m6 processes to be adjusted from other processes, the determined m6 processes to be adjusted are all allocated process throughput for the kth target outbound call model, k is an integer between [1, n2], and m6 is less than or equal to m1.

[0424] S1502, determining throughputs to be adjusted corresponding to m6 processes to be adjusted respectively.

[0425] In the above S1502, for the qth process to be adjusted among the m6 processes to be adjusted, q is an integer between [1, m6], and the throughput to be adjusted corresponding to the qth process to be adjusted can be determined by the following method: select the minimum value from the difference between the minimum throughput of the process and the allocated throughput corresponding to the m1th process, the difference between the allocated throughput corresponding to the qth process to be adjusted and the minimum throughput of the process, and the process throughput allocated by the qth process to be adjusted for the kth target outbound call model, as the throughput to be adjusted corresponding to the qth process to be adjusted.

[0426] In addition, the allocated throughput corresponding to the qth process to be adjusted may also be updated. Specifically, the allocated throughput corresponding to the qth process to be adjusted minus the throughput to be adjusted corresponding to the qth process to be adjusted is the updated allocated throughput corresponding to the qth process to be adjusted.

[0427] In addition, the target outbound call model to be loaded by the qth process to be adjusted may be updated, and the process throughput allocated by the qth process to be adjusted to each target outbound call model may be updated.

[0428] S1503: Update the allocated throughput corresponding to the m1th process according to the throughputs to be adjusted corresponding to the m6 processes to be adjusted.

[0429] In the above S1503, the allocated throughput corresponding to the m1th process is added to the throughputs to be adjusted corresponding to the m6 processes to be adjusted, which is the allocated throughput corresponding to the m1th process after update.

[0430] Fig.15 A first loading strategy adjustment method is provided.

[0431] For example, the first loading strategy is shown in Table 10. The allocated throughput corresponding to process 13 is 314. Since the allocated throughput corresponding to process 13 (i.e., 314) is less than the minimum process throughput 372, and the allocated throughput corresponding to process 13 (i.e., 314) is greater than the ratio of the expected process throughput to the second preset value (i.e., 200=400 / 2), the following steps are performed:

[0432] Determine the three target outbound call models for allocating process throughput to process 13, namely, outbound call model 2, outbound call model 4, and outbound call model 7.

[0433] For outbound call model 2, it can be determined that the five processes to be adjusted for allocating process throughput for outbound call model 2 are process 1, process 2, process 6, process 10, and process 12, respectively, as shown in Table 11.

[0434] Table 11.

[0435]

[0436] For process 1, it can be determined that the throughput to be adjusted for process 1 is 56 (i.e., 56=min(372-314,428-372,172)). Therefore, the process throughput allocated by process 1 for outbound call model 2 is adjusted to 116 (i.e., 116=172-56), and the process throughput allocated by process 13 for outbound call model 2 is adjusted to 112 (i.e., 112=56+56). The currently allocated throughput for process 13 is 370. Since the allocated throughput for process 13 (i.e., 370) is less than the minimum process throughput 372, the following steps are performed:

[0437] For process 2, it can be determined that the throughput to be adjusted corresponding to process 2 is 2 (i.e., 2 = min (372-370, 428-372, 172)). Therefore, the process throughput allocated by process 2 for outbound call model 2 is adjusted to 170 (i.e., 170 = 172-2), and the process throughput allocated by process 13 for outbound call model 2 is adjusted to 114 (i.e., 114 = 112 + 2). The currently allocated throughput corresponding to process 13 is 372. Since the allocated throughput corresponding to process 13 (i.e., 372) is equal to the minimum process throughput 372, the iteration stops.

[0438] Finally, the second loading strategy can be determined, which includes the number of processes to be generated 13, multiple target outbound call models to be loaded by each process, and the process throughput allocated to each process for the multiple target outbound call models, as shown in Table 12.

[0439] Table 12.

[0440]

[0441] In one possible implementation, by Fig.13 After the second loading strategy is determined in the steps shown, the second loading strategy can also be executed. Specifically, according to the second loading strategy, m2 processes are generated, and n3 target outbound call models are loaded for each of the m2 processes, and corresponding process throughputs are respectively allocated to the n3 target outbound call models loaded for each of the m2 processes.

[0442] It should be understood that the execution method of the second loading strategy can refer to the execution method of the first loading strategy, which will not be elaborated here.

[0443] In one possible implementation, after executing the first loading strategy or the second loading strategy, the actual model throughput of each outbound call model can be monitored. The reason for monitoring the actual model throughput of each outbound call model is that the actual model throughput of each outbound call model is determined based on the actual situation of each outbound call model in dialing an outbound number. As time goes by, the actual model throughput of each outbound call model will gradually deviate from the assigned process throughput corresponding to each outbound call model, resulting in a large deviation from the expected outbound call model assigned to the process when the process actually loads the outbound call model. In this regard, the present application proposes to adjust the first loading strategy or the second loading strategy according to the actual model throughput of each outbound call model. To facilitate subsequent discussion, the following text will take the execution of the first loading strategy as an example for detailed description, including the following. Fig.16 The following steps are shown.

[0444] S1601, monitor m1 processes according to the monitoring window, determine the actual model throughputs corresponding to m4 processes and n4 target outbound call models loaded by each of the m4 processes; the throughput of each of the m4 processes is greater than 0; m4 is less than or equal to m1, and n4 is an integer greater than 0.

[0445] In the embodiment of the present application, in order to enable timely adjustment of the first loading strategy, the length of the monitoring window is set to be smaller than the length of the current outbound call window.

[0446] In the embodiment of the present application, when the outbound call model actually dials the outbound call number, the answering rate and average number of interactions corresponding to the monitoring window can be determined according to whether each outbound call number in the monitoring window is answered and the number of interactions of each outbound call number. The answering rate and average number of interactions corresponding to the monitoring window will affect the actual throughput of the model corresponding to the outbound call model.

[0447] S1602: According to the actual model throughputs corresponding to the n4 target outbound call models loaded by each of the m4 processes, the first loading strategy is adjusted to obtain a third loading strategy.

[0448] In the embodiment of the present application, the third loading strategy includes the number of processes m4 to be generated, the n4 target outbound call models to be loaded by each of the m4 processes, and the process throughput allocated to the n4 target outbound call models by each of the m4 processes. Among them, the process throughput allocated to the n4 target outbound call models by each of the m4 processes is the actual model throughput corresponding to the n4 target outbound call models loaded by each of the m4 processes.

[0449] In the above S1602, for the sth process among the m4 processes, s is an integer between [1, m4], execute the following Fig.17The following steps are shown:

[0450] S1701, determine n4 target outbound call models to be loaded by the sth process.

[0451] S1702: Determine the number of reusable outbound call models corresponding to each of the m1 processes according to the n4 target outbound call models to be loaded by the sth process.

[0452] In the embodiment of the present application, for one process among the m1 processes, the reusable outbound call model corresponding to the process is the target outbound call model among the n4 target outbound call models loaded into the process.

[0453] S1703: Select the process corresponding to the largest number of reusable outbound call models as the reusable process.

[0454] S1704: According to the n4 target outbound call models loaded by the sth process, adjust the target outbound call model to be loaded by the reusable process.

[0455] In the embodiment of the present application, if the reusable process has loaded a target outbound call model other than the above n4 target outbound call models, it is unloaded; if the reusable process has not loaded any of the above n4 target outbound call models, it is loaded, so that the reusable process finally loads the above n4 target outbound call models.

[0456] S1705: According to the actual model throughputs respectively corresponding to the n4 target outbound call models loaded by the sth process, adjust the process throughputs respectively allocated by the reusable process to the n4 target outbound call models.

[0457] In the above S1705, the actual model throughputs corresponding to the n4 target outbound call models loaded by the sth process are used as the process throughputs respectively allocated to the n4 target outbound call models by the reusable process for adjusting.

[0458] exist Fig.17 In the method shown, the process corresponding to the largest number of reusable outbound call models is selected as the reusable process corresponding to the sth process, rather than selecting the process with the same process ID (i.e., process number) as the reusable process corresponding to the sth process from the m1 processes. The main reason is that through Fig.17 The reusable process corresponding to the s-th process determined by the method shown has already loaded part or all of the target outbound call models to be loaded by the s-th process. Therefore, the number of target outbound call models to be loaded by the reusable process corresponding to the s-th process can be minimized, which is conducive to improving the generation efficiency of the third strategy.

[0459] For each of the m4 processes, execute the following Fig.17The third strategy can be obtained by following the steps shown.

[0460] For example, the first loading strategy is shown in Table 10. After the first loading strategy is executed, according to the monitoring window, the 11 processes monitored (the process numbers of the 11 processes are set to A1 to A11 in sequence) and the actual model throughputs corresponding to the target outbound call models loaded by each of the 11 processes are shown in Table 13.

[0461] Table 13.

[0462]

[0463]

[0464] It should be understood that in order to distinguish the process numbers in Table 10 and avoid confusion with the process numbers in Table 10, the process numbers in Table 13 are recorded as A1 to A11. Moreover, the process numbers A1 to A11 do not correspond to the process numbers 1 to 13 in sequence. For example, process A1 and process 1 may be the same process or different processes.

[0465] For process A1 , it is determined that the target outbound call models to be loaded by process A1 include outbound call model 1 , outbound call model 2 , and outbound call model 10 .

[0466] According to the target outbound call model to be loaded by process A1, the number of reusable outbound call models corresponding to each process from process 1 to process 13 in Table 1 is determined, as shown in Table 14.

[0467] Table 14.

[0468] Program number The number of reusable outbound call models 1 3 2 3 3 0 4 1 5 0 6 1 7 0 8 1 9 0 10 1 11 1 12 2 13 1

[0469] Since the number of reusable outbound call models corresponding to process 1 and process 2 is the largest, both 3, one of the two processes can be selected as the reusable process. In the embodiment of the present application, process 1 is selected as the reusable process.

[0470] Since the target outbound call model to be loaded by process A1 is the same as the target outbound call model to be loaded by process 1, both including outbound call model 1, outbound call model 2 and outbound call model 10, there is no need to adjust the target outbound call model to be loaded by process 1.

[0471] Since the actual model throughputs corresponding to the outbound call model 1, outbound call model 2 and outbound call model 10 loaded by process A1 are 200, 138 and 112 respectively, the process throughputs corresponding to the outbound call model 1, outbound call model 2 and outbound call model 10 loaded by process 1 are set to 200, 138 and 112.

[0472] It should be understood that the processing methods of processes A2 to A11 refer to process A1 and are not described in detail here.

[0473] In one possible implementation, by Fig.16 After the third loading strategy is determined in the steps shown, the third loading strategy can also be executed. Specifically, according to the third loading strategy, m4 processes are generated, and n4 target outbound call models are loaded for each of the m4 processes, and corresponding process throughputs are respectively allocated to the n4 target outbound call models loaded for each of the m4 processes.

[0474] It should be understood that the execution method of the third loading strategy can refer to the execution method of the first loading strategy, which will not be elaborated here.

[0475] Based on the same technical concept, the embodiment of the present application provides a model loading strategy determination device, such as Fig.18 As shown, the model loading strategy determination device 1800 includes:

[0476] The acquisition module 1801 is used to acquire outbound call files corresponding to n1 outbound call scenarios, wherein the outbound call files include the outbound call numbers to be dialed; n1 is an integer greater than 0;

[0477] The determination module 1802 is used to determine the outbound call models corresponding to the n1 outbound call scenarios respectively, and determine the model estimated throughput corresponding to the n1 outbound call models respectively according to the outbound call numbers in the n1 outbound call files;

[0478] The determination module 1802 is also used to determine a first loading strategy based on the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that a process can provide, wherein the first loading strategy includes the number of processes m1 to be generated, the n2 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each of the m1 processes; the expected process throughput is used to identify that the load of the process is in an optimal state; the n1 outbound call models include the n2 target outbound call models; n2 is less than or equal to n1, and m1 and n2 are both integers greater than 0.

[0479] Optionally, the determining module 1802 is specifically configured to:

[0480] Determine a current outbound call window and a historical outbound call window corresponding to the current outbound call window;

[0481] From the i-th outbound call file, determine the outbound call number within the current outbound call window; i is an integer between [1, n1];

[0482] Determine the historical outbound call number within the historical outbound call window from the historical outbound call file corresponding to the i-th outbound call file;

[0483] The model estimated throughput corresponding to the i-th outbound call model is determined according to the outbound call number in the current outbound call window and the historical outbound call number in the historical outbound call window.

[0484] Optionally, the determining module 1802 is specifically configured to:

[0485] Determine the number of outbound call numbers corresponding to the current outbound call window according to the outbound call numbers in the current outbound call window;

[0486] Determine the historical answering rate and the historical average number of interactions corresponding to the historical outbound call window according to the historical outbound call number in the historical outbound call window;

[0487] The model estimated throughput corresponding to the i-th outbound call model is determined according to the number of outbound call numbers corresponding to the current outbound call window, the historical answering rate corresponding to the historical outbound call window, and the historical average number of interactions.

[0488] Optionally, the determining module 1802 is specifically configured to:

[0489] Determine the model residual throughputs respectively corresponding to the n1 outbound call models according to the model estimated throughputs respectively corresponding to the n1 outbound call models;

[0490] Repeat the following process until the model residual throughputs corresponding to the n1 outbound call models are 0:

[0491] Identify a process;

[0492] Selecting n2 target outbound call models from the n1 outbound call models, and determining the process throughputs respectively allocated by the one process to the n2 target outbound call models according to the expected process throughputs;

[0493] The process throughputs respectively allocated by the one process to the n2 target outbound call models are used to update the model remaining throughputs respectively corresponding to the n1 outbound call models.

[0494] Optionally, the determining module 1802 is specifically configured to:

[0495] Determine an allocated throughput corresponding to the one process, and initialize the allocated throughput corresponding to the one process;

[0496] Repeat the following steps until the allocated throughput corresponding to the one process is within a process throughput range, where the process throughput range is determined according to the expected throughput of the process:

[0497] According to the sorting rule, the model remaining throughputs corresponding to the n1 outbound call models are sorted to obtain a first sequence;

[0498] Determine a target outbound call model according to a model remaining throughput corresponding to the outbound call model located at the first position in the first sequence;

[0499] Determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model;

[0500] The process throughput allocated by the one process to the target outbound call model is used to update the allocated throughput corresponding to the one process, and the model remaining throughputs respectively corresponding to the n1 outbound call models are updated.

[0501] Optionally, the determining module 1802 is specifically configured to:

[0502] Determine the number S of determined outbound call models, where the determined outbound call model is a target outbound call model corresponding to the allocated throughput corresponding to the one process;

[0503] In the case that the number S of the determined outbound call models is an even number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in descending order to obtain a first sequence;

[0504] In the case that the number S of the determined outbound call models is an odd number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in ascending order to obtain a first sequence.

[0505] Optionally, the determining module 1802 is specifically configured to:

[0506] Determining a model maximum throughput based on the expected throughput of the process;

[0507] Determine the model minimum throughput according to the process expected throughput and the model estimated throughputs corresponding to the n1 outbound call models respectively;

[0508] The process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput.

[0509] Optionally, the determining module 1802 is specifically configured to:

[0510] In a case where the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than or equal to a first preset value, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput or the model minimum throughput corresponding to the target outbound call model;

[0511] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, the process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model and the model maximum throughput.

[0512] Optionally, the determining module 1802 is specifically configured to:

[0513] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than a first preset value, using the model remaining throughput corresponding to the target outbound call model as the process throughput allocated by the one process to the target outbound call model;

[0514] When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is equal to a first preset value, the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0515] Optionally, the determining module 1802 is specifically configured to:

[0516] Determining a minimum throughput of the process based on the expected throughput of the process;

[0517] Determine the throughput to be allocated corresponding to the process according to the minimum throughput of the process and the allocated throughput corresponding to the process;

[0518] In the case where the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, if the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the throughput to be allocated corresponding to the one process, and the model minimum throughput;

[0519] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, the process throughput allocated to the target outbound call model by the one process is determined according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput.

[0520] Optionally, the determining module 1802 is specifically configured to:

[0521] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then the process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model and the model minimum throughput; the process throughput control value is determined according to the process throughput range;

[0522] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the process throughput allocated by the one process to the target outbound call model is determined according to the throughput to be allocated corresponding to the one process and the model minimum throughput.

[0523] Optionally, the determining module 1802 is specifically configured to:

[0524] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is not equal to the model estimated throughput corresponding to the target outbound call model, then the model remaining throughput corresponding to the target outbound call model is used as the process throughput allocated by the one process to the target outbound call model;

[0525] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is equal to the model estimated throughput corresponding to the target outbound call model, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0526] Optionally, the determining module 1802 is specifically configured to:

[0527] If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the maximum value of the throughput to be allocated corresponding to the one process and the model minimum throughput is selected as the process throughput allocated by the one process to the target outbound call model.

[0528] Optionally, the determining module 1802 is specifically configured to:

[0529] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, then determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model minimum throughput, and the to-be-allocated throughput corresponding to the one process;

[0530] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, then the process throughput allocated to the target outbound call model by the process is determined according to the model minimum throughput, the model maximum throughput, and the throughput to be allocated corresponding to the one process.

[0531] Optionally, the determining module 1802 is specifically configured to:

[0532] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is less than the to-be-allocated throughput corresponding to the one process, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model;

[0533] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is greater than or equal to the throughput to be allocated corresponding to the one process, then the maximum value between the throughput to be allocated corresponding to the one process and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

[0534] Optionally, the determining module 1802 is specifically configured to:

[0535] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then the model maximum throughput is used as the process throughput allocated by the one process to the target outbound call model; the process throughput control value is determined according to the process throughput range;

[0536] If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the sum of the throughput to be allocated corresponding to the one process and the process throughput control value is used as the process throughput allocated by the one process to the target outbound call model.

[0537] Optionally, an adjustment module 1803 is further included, and the adjustment module 1803 is specifically used to:

[0538] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum process throughput, the first loading strategy is adjusted to obtain the second loading strategy; the second loading strategy includes the number of processes m2 to be generated, the n3 target outbound call models to be loaded by each of the m2 processes, and the process throughput allocated by each of the m2 processes to the n3 target outbound call models; the minimum process throughput is determined based on the expected throughput of the process, m2 is less than or equal to m1, and n3 is an integer greater than 0.

[0539] Optionally, the adjustment module 1803 is specifically configured to:

[0540] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is less than the ratio of the expected throughput of the process to the second preset value, the allocated throughput corresponding to the m1th process is allocated to other processes to obtain the second loading strategy; the other processes are processes other than the m1th process among the m1 processes; the second loading strategy includes the number of processes to be generated (m1)-1, the n3 target outbound call models to be loaded by each of the (m1)-1 processes, and the process throughput allocated by each of the (m1)-1 processes to the n3 target outbound call models;

[0541] If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is greater than or equal to the ratio of the expected throughput of the process to the second preset value, the allocated throughputs corresponding to the other processes are allocated to the m1th process to obtain the second loading strategy; the second loading strategy includes the number of processes m1 to be generated, n3 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated by each of the m1 processes to the n3 target outbound call models.

[0542] Optionally, an execution module 1804 is further included, and the execution module 1804 is specifically used to:

[0543] According to the first loading strategy, m1 processes are generated, and n2 target outbound call models are loaded for each of the m1 processes, and corresponding process throughputs are respectively allocated to the n2 target outbound call models loaded for each of the m1 processes.

[0544] Optionally, the execution module 1804 is further configured to:

[0545] Obtain an outbound call number in the outbound call file, and generate a random number for the outbound call number;

[0546] Determine a target outbound call model corresponding to the one outbound call number, and determine m3 processes where the target outbound call model corresponding to the one outbound call number is located; m3 is less than or equal to m1;

[0547] According to the random number, a target process is selected from the m3 processes, and the one outbound call number is allocated to the target process, so that the target outbound call model corresponding to the one outbound call number deployed in the target process dials the one outbound call number.

[0548] Optionally, the execution module 1804 is specifically configured to:

[0549] Determine the process throughputs allocated by the m3 processes to the target outbound call model corresponding to the one outbound call number;

[0550] Determine the value intervals corresponding to the m3 processes respectively according to the process throughputs respectively allocated by the m3 processes to the target outbound call model corresponding to the one outbound call number;

[0551] A process corresponding to the value interval of the random number is selected as the target process.

[0552] Optionally, the adjustment module 1803 is further configured to:

[0553] According to the monitoring window, the m1 processes are monitored to determine the actual model throughputs corresponding to the m4 processes and the n4 target outbound call models loaded by each of the m4 processes; the throughput of each of the m4 processes is greater than 0; m4 is less than or equal to m1, and n4 is an integer greater than 0;

[0554] According to the actual model throughputs corresponding to the n4 target outbound call models loaded by each of the m4 processes, the first loading strategy is adjusted to obtain a third loading strategy; the third loading strategy includes the number of processes m4 to be generated, the n4 target outbound call models to be loaded by each of the m4 processes, and the process throughput allocated by each of the m4 processes to the n4 target outbound call models.

[0555] Based on the same technical concept, the embodiment of the present application provides a computer device, which may be a terminal or a server, such as Fig.19 As shown, it includes at least one processor 1901 and a memory 1902 connected to the at least one processor. The specific connection medium between the processor 1901 and the memory 1902 is not limited in the embodiment of the present application. Fig.19 For example, the processor 1901 and the memory 1902 are connected via a bus. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0556] In an embodiment of the present application, the memory 1902 stores instructions that can be executed by at least one processor 1901. The at least one processor 1901 can execute the steps included in the above-mentioned model loading strategy determination method by executing the instructions stored in the memory 1902.

[0557] Among them, the processor 1901 is the control center of the computer device, and various interfaces and lines can be used to connect various parts of the computer device, and the model loading strategy can be determined by running or executing instructions stored in the memory 1902 and calling data stored in the memory 1902. Optionally, the processor 1901 may include one or more processing units, and the processor 1901 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 1901. In some embodiments, the processor 1901 and the memory 1902 may be implemented on the same chip, and in some embodiments, they may also be implemented separately on independent chips.

[0558] Processor 1901 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware processor for execution, or can be executed by a combination of hardware and software modules in the processor.

[0559] Memory 1902, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules. Memory 1902 may include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory, random access memory (Random Access Memory, RAM), static random access memory (Static Random Access Memory, SRAM), programmable read-only memory (Programmable Read Only Memory, PROM), read-only memory (Read Only Memory, ROM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), magnetic memory, disk, optical disk, etc. Memory 1902 is any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 1902 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0560] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program that can be executed by a computer device. When the program runs on the computer device, the computer device executes the steps of the above-mentioned model loading strategy determination method.

[0561] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0562] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0563] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0564] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0565] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for determining a model loading strategy, characterized in that: include: Obtain outbound call files corresponding to n1 outbound call scenarios, wherein the outbound call files include the outbound call number to be dialed; n1 is an integer greater than 0; Determine the outbound call models corresponding to the n1 outbound call scenarios, and determine the model estimated throughput corresponding to the n1 outbound call models according to the outbound call numbers in the n1 outbound call files; A first loading strategy is determined based on the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that a process can provide. The first loading strategy includes the number of processes m1 to be generated, the n2 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each of the m1 processes; the expected process throughput is used to identify that the load of the process is in an optimal state; the n1 outbound call models include the n2 target outbound call models; n2 is less than or equal to n1, and m1 and n2 are both integers greater than 0.

2. The method according to claim 1, characterized in that The step of determining the model estimated throughputs corresponding to the n1 outbound call models respectively according to the outbound call numbers in the n1 outbound call files includes: Determine a current outbound call window and a historical outbound call window corresponding to the current outbound call window; From the i-th outbound call file, determine the outbound call number within the current outbound call window; i is an integer between [1, n1]; Determine the historical outbound call number within the historical outbound call window from the historical outbound call file corresponding to the i-th outbound call file; The model estimated throughput corresponding to the i-th outbound call model is determined according to the outbound call number in the current outbound call window and the historical outbound call number in the historical outbound call window.

3. The method according to claim 2, characterized in that The determining, according to the outbound call number in the current outbound call window and the historical outbound call number in the historical outbound call window, a model estimated throughput corresponding to the i-th outbound call model includes: Determine the number of outbound call numbers corresponding to the current outbound call window according to the outbound call numbers in the current outbound call window; Determine the historical answering rate and the historical average number of interactions corresponding to the historical outbound call window according to the historical outbound call number in the historical outbound call window; The model estimated throughput corresponding to the i-th outbound call model is determined according to the number of outbound call numbers corresponding to the current outbound call window, the historical answering rate corresponding to the historical outbound call window, and the historical average number of interactions.

4. The method according to claim 1, characterized in that The determining of the first loading strategy according to the model estimated throughputs respectively corresponding to the n1 outbound call models and the process expected throughput that can be provided by one process includes: Determine the model residual throughputs respectively corresponding to the n1 outbound call models according to the model estimated throughputs respectively corresponding to the n1 outbound call models; Repeat the following process until the model residual throughputs corresponding to the n1 outbound call models are 0: Identify a process; Selecting n2 target outbound call models from the n1 outbound call models, and determining the process throughputs respectively allocated by the one process to the n2 target outbound call models according to the expected process throughputs; The process throughputs respectively allocated by the one process to the n2 target outbound call models are used to update the model remaining throughputs respectively corresponding to the n1 outbound call models.

5. The method according to claim 4, characterized in that The selecting n2 target outbound call models from the n1 outbound call models, and determining the process throughputs respectively allocated by the one process to the n2 target outbound call models according to the expected process throughputs, and updating the model residual throughputs respectively corresponding to the n1 outbound call models by using the process throughputs respectively allocated by the one process to the n2 target outbound call models, comprises: Determine an allocated throughput corresponding to the one process, and initialize the allocated throughput corresponding to the one process; Repeat the following steps until the allocated throughput corresponding to the one process is within a process throughput range, where the process throughput range is determined according to the expected throughput of the process: According to the sorting rule, the model remaining throughputs corresponding to the n1 outbound call models are sorted to obtain a first sequence; Determine a target outbound call model according to a model remaining throughput corresponding to the outbound call model located at the first position in the first sequence; Determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model; The process throughput allocated by the one process to the target outbound call model is used to update the allocated throughput corresponding to the one process, and the model remaining throughputs respectively corresponding to the n1 outbound call models are updated.

6. The method according to claim 5, characterized in that The method of sorting the model remaining throughputs respectively corresponding to the n1 outbound call models according to the sorting rule to obtain a first sequence includes: Determine the number S of determined outbound call models, where the determined outbound call model is a target outbound call model corresponding to the allocated throughput corresponding to the one process; In the case that the number S of the determined outbound call models is an even number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in descending order to obtain a first sequence; In the case that the number S of the determined outbound call models is an odd number, the model remaining throughputs corresponding to the n1 outbound call models are sorted in ascending order to obtain a first sequence.

7. The method according to claim 5, characterized in that The determining, according to the model remaining throughput corresponding to the target outbound call model, the process throughput allocated by the one process to the target outbound call model comprises: Determining a model maximum throughput based on the expected throughput of the process; Determine the model minimum throughput according to the process expected throughput and the model estimated throughputs corresponding to the n1 outbound call models respectively; The process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput.

8. The method according to claim 7, characterized in that The determining, according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput, the process throughput allocated by the one process to the target outbound call model comprises: In a case where the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than or equal to a first preset value, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput or the model minimum throughput corresponding to the target outbound call model; When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, the process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model and the model maximum throughput.

9. The method according to claim 8, characterized in that The determining, when a ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than or equal to a first preset value, according to the model remaining throughput or the model minimum throughput corresponding to the target outbound call model, the process throughput allocated by the one process to the target outbound call model includes: When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is less than a first preset value, using the model remaining throughput corresponding to the target outbound call model as the process throughput allocated by the one process to the target outbound call model; When the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is equal to a first preset value, the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

10. The method according to claim 8, characterized in that The method further comprises: determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model and the model maximum throughput when the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, comprising: Determining a minimum throughput of the process based on the expected throughput of the process; Determine the throughput to be allocated corresponding to the process according to the minimum throughput of the process and the allocated throughput corresponding to the process; In the case where the ratio of the model remaining throughput corresponding to the target outbound call model to the model minimum throughput is greater than a first preset value, if the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the throughput to be allocated corresponding to the one process, and the model minimum throughput; If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, the process throughput allocated to the target outbound call model by the one process is determined according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput.

11. The method according to claim 10, characterized in that If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the throughput to be allocated corresponding to the one process, and the model minimum throughput, including: If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then the process throughput allocated by the one process to the target outbound call model is determined according to the model remaining throughput corresponding to the target outbound call model and the model minimum throughput; the process throughput control value is determined according to the process throughput range; If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the process throughput allocated by the one process to the target outbound call model is determined according to the throughput to be allocated corresponding to the one process and the model minimum throughput.

12. The method according to claim 11, characterized in that If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model and the model minimum throughput, including: If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is not equal to the model estimated throughput corresponding to the target outbound call model, then the model remaining throughput corresponding to the target outbound call model is used as the process throughput allocated by the one process to the target outbound call model; If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, and the model remaining throughput corresponding to the target outbound call model is equal to the model estimated throughput corresponding to the target outbound call model, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

13. The method according to claim 11, characterized in that If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, determining the process throughput allocated by the one process to the target outbound call model according to the to-be-allocated throughput corresponding to the one process and the model minimum throughput, including: If the model remaining throughput corresponding to the target outbound call model is less than or equal to the model maximum throughput, and the model remaining throughput corresponding to the target outbound call model is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the maximum value of the throughput to be allocated corresponding to the one process and the model minimum throughput is selected as the process throughput allocated by the one process to the target outbound call model.

14. The method according to claim 10, characterized in that If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model maximum throughput, and the model minimum throughput, including: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, then determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model minimum throughput, and the to-be-allocated throughput corresponding to the one process; If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, then the process throughput allocated to the target outbound call model by the process is determined according to the model minimum throughput, the model maximum throughput, and the throughput to be allocated corresponding to the one process.

15. The method according to claim 14, characterized in that If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model remaining throughput corresponding to the target outbound call model, the model minimum throughput, and the to-be-allocated throughput corresponding to the one process, including: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is less than the to-be-allocated throughput corresponding to the one process, then the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model; If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is less than the model minimum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model minimum throughput is greater than or equal to the throughput to be allocated corresponding to the one process, then the maximum value between the throughput to be allocated corresponding to the one process and the model minimum throughput is used as the process throughput allocated by the one process to the target outbound call model.

16. The method according to claim 14, characterized in that If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, determining the process throughput allocated by the one process to the target outbound call model according to the model minimum throughput, the model maximum throughput, and the to-be-allocated throughput corresponding to the one process, including: If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is less than or equal to the sum of the to-be-allocated throughput corresponding to the one process and the process throughput control value, then the model maximum throughput is used as the process throughput allocated by the one process to the target outbound call model; the process throughput control value is determined according to the process throughput range; If the model remaining throughput corresponding to the target outbound call model is greater than the model maximum throughput, and the difference between the model remaining throughput corresponding to the target outbound call model and the model maximum throughput is greater than or equal to the model minimum throughput, and the model maximum throughput is greater than the sum of the throughput to be allocated corresponding to the one process and the process throughput control value, then the sum of the throughput to be allocated corresponding to the one process and the process throughput control value is used as the process throughput allocated by the one process to the target outbound call model.

17. The method according to claim 1, wherein: Also includes: If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, adjusting the first loading strategy to obtain a second loading strategy; The second loading strategy includes the number of processes m2 to be generated, the n3 target outbound call models to be loaded by each of the m2 processes, and the process throughput allocated to the n3 target outbound call models by each of the m2 processes; the minimum process throughput is determined based on the expected throughput of the process, m2 is less than or equal to m1, and n3 is an integer greater than 0.

18. The method according to claim 17, characterized in that If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, adjusting the first loading strategy to obtain the second loading strategy includes: If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is less than the ratio of the expected throughput of the process to the second preset value, the allocated throughput corresponding to the m1th process is allocated to other processes to obtain the second loading strategy; the other processes are processes other than the m1th process among the m1 processes; the second loading strategy includes the number of processes to be generated (m1)-1, the n3 target outbound call models to be loaded by each of the (m1)-1 processes, and the process throughput allocated by each of the (m1)-1 processes to the n3 target outbound call models; If the allocated throughput corresponding to the m1th process among the m1 processes is less than the minimum throughput of the process, and the allocated throughput corresponding to the m1th process is greater than or equal to the ratio of the expected throughput of the process to the second preset value, the allocated throughputs corresponding to the other processes are allocated to the m1th process to obtain the second loading strategy; the second loading strategy includes the number of processes m1 to be generated, n3 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated by each of the m1 processes to the n3 target outbound call models.

19. The method according to any one of claims 1 to 18, characterized in that: Also includes: According to the first loading strategy, m1 processes are generated, and n2 target outbound call models are loaded for each of the m1 processes, and corresponding process throughputs are respectively allocated to the n2 target outbound call models loaded for each of the m1 processes.

20. The method of claim 19, wherein: Also includes: Obtain an outbound call number in the outbound call file, and generate a random number for the outbound call number; Determine a target outbound call model corresponding to the outbound call number, and determine m3 processes where the target outbound call model corresponding to the outbound call number is located; m3 is less than or equal to m1; According to the random number, a target process is selected from the m3 processes, and the one outbound call number is allocated to the target process, so that the target outbound call model corresponding to the one outbound call number deployed in the target process dials the one outbound call number.

21. The method of claim 20, wherein: The step of selecting a target process from the m3 processes according to the random number includes: Determine the process throughputs allocated by the m3 processes to the target outbound call model corresponding to the one outbound call number; Determine the value intervals corresponding to the m3 processes respectively according to the process throughputs allocated by the m3 processes to the target outbound call model corresponding to the one outbound call number; A process corresponding to the value interval of the random number is selected as the target process.

22. The method of claim 20, wherein: Also includes: According to the monitoring window, the m1 processes are monitored to determine the actual model throughputs corresponding to the m4 processes and the n4 target outbound call models loaded by each of the m4 processes; the throughput of each of the m4 processes is greater than 0; m4 is less than or equal to m1, and n4 is an integer greater than 0; According to the actual model throughputs respectively corresponding to the n4 target outbound call models loaded by each process in the m4 processes, the first loading strategy is adjusted to obtain a third loading strategy; The third loading strategy includes the number of processes m4 to be generated, the n4 target outbound call models to be loaded by each of the m4 processes, and the process throughput allocated to the n4 target outbound call models by each of the m4 processes.

23. A model loading strategy determination device, characterized in that: include: An acquisition module, used to acquire outbound call files corresponding to n1 outbound call scenarios, wherein the outbound call files include an outbound call number to be dialed; n1 is an integer greater than 0; A determination module, used to determine the outbound call models corresponding to n1 outbound call scenarios respectively, and determine the model estimated throughput corresponding to the n1 outbound call models respectively according to the outbound call numbers in the n1 outbound call files; The determination module is further used to determine a first loading strategy based on the model estimated throughputs corresponding to the n1 outbound call models and the expected process throughput that a process can provide, wherein the first loading strategy includes the number of processes m1 to be generated, the n2 target outbound call models to be loaded by each of the m1 processes, and the process throughput allocated to the n2 target outbound call models by each of the m1 processes; the expected process throughput is used to identify that the load of the process is in an optimal state; the n1 outbound call models include the n2 target outbound call models; n2 is less than or equal to n1, and m1 and n2 are both integers greater than 0.

24. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of any one of the methods of claims 1 to 22 are implemented.

25. A computer-readable storage medium, characterized in that: It stores a computer program executable by a computer device. When the program is run on the computer device, the computer device executes the steps of any method described in claims 1 to 22.

Images