Degradation migration method and device for shopping guide dialogue service, equipment and storage medium
By adopting the downgrade and migration method of shopping guide dialogue services in e-commerce platforms, and using user experience weighted scores and principal component analysis methods to optimize the deployment of shopping guide dialogue services, the problem of server overload on e-commerce platforms during the big promotion period is solved, ensuring user experience stability.
Patent Information
- Application Number
- CN202411200588.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-11-15
AI Technical Summary
During the big promotion period, e-commerce platforms overloaded servers and increased GPU resources and traffic suddenly, making it difficult to meet the experience needs of shopping guide services through traditional expansion methods, resulting in a decline in user experience.
By applying the downgrade migration method of shopping guide dialogue service in the e-commerce service system, using the sum of user experience weighted scores to maximize the optimization goal, the optimization migration of shopping guide dialogue service is carried out according to the resource limitations of each server, and the user is calculated using the principal component analysis method. Experience weight, optimize the deployment of shopping guide conversation services on each server to keep the user experience from decreasing.
In the face of limited resources and sudden traffic, by optimizing the deployment of shopping guide dialogue services, the smooth provision of shopping guide services is achieved, the stability of the user experience is maintained, and the decline in service experience is avoided.
Smart Images

Figure CN118964040A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of e-commerce technology, and more specifically, to a method, device, equipment and storage medium for downgrading and migrating a shopping guide dialogue service. Background Art
[0002] With the continuous development of e-commerce technology, e-commerce platforms seek to recommend product information that better meets user needs to provide users with a better online shopping experience. When users shop online, they can quickly find related products of interest through shopping guide assistants, which are equivalent to intelligent customer service and provide users with real-time shopping guide services. Excellent shopping guide services can bring users a good user experience, which is an important guarantee for the development of the industry.
[0003] However, e-commerce systems have limited resources. During big sales events, a large influx of users consumes a lot of GPU resources, causing a sudden increase in traffic and overloading of service machines. It is difficult to meet production needs through traditional capacity expansion methods, resulting in a failure to meet users' experience needs for shopping guide services.
[0004] Based on this, how to maintain the user experience of shopping guide services without decreasing between the priority GPU resources and the sudden increase in traffic is an issue that needs attention. Summary of the invention
[0005] In view of the above problems, the present application provides a method, device, equipment and storage medium for downgrading and migrating a shopping guide dialogue service, so as to maintain the user experience of the shopping guide service without degradation between the priority GPU resources and the sudden increase in traffic.
[0006] In order to achieve the above objectives, the specific plan is proposed as follows:
[0007] A method for downgrading and migrating a shopping guide dialogue service is applied to an e-commerce service system, wherein the e-commerce service system includes a plurality of servers, and at least one shopping guide dialogue service is deployed in each server. The method includes:
[0008] When there is a server overload in the e-commerce service system, the optimization goal is to maximize the sum of the weighted scores of user experience of various shopping guide dialogue services of the e-commerce service system, and the current resource limit of each server is used as the optimization constraint condition to determine the optimization migration result of the shopping guide dialogue service deployed in each server;
[0009] Based on the optimization migration result, the shopping guide dialogue service deployed in each server is migrated;
[0010] The calculation process of the user experience weighted score of each shopping guide dialogue service of the e-commerce service system is as follows:
[0011] Determine a user experience score for each shopping guide dialogue service of the e-commerce service system;
[0012] Calculate the user experience weight of each shopping guide dialogue service based on the deployment of various shopping guide dialogue services in various servers;
[0013] The user experience score of each shopping guide dialogue service is multiplied by the user experience weight of the shopping guide dialogue service to obtain a weighted user experience score of the shopping guide dialogue service.
[0014] Optionally, the optimization objective is:
[0015]
[0016] in, The weighted score of the user experience of the i-th shopping guide dialogue service;
[0017] The optimization constraints are:
[0018]
[0019] in, For each server under the current resource constraints, Serving the i-th shopping guide dialogue, is the GPU resource occupied by the i-th shopping guide dialogue service on a single server, is the number of servers deployed with the i-th shopping guide dialogue service.
[0020] Optionally, the calculating the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers includes:
[0021] Through principal component analysis, the user experience weight of each shopping guide dialogue service is calculated according to the deployment of various shopping guide dialogue services in various servers.
[0022] Optionally, the calculating the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers includes:
[0023] The number of servers deployed with each shopping guide dialogue service is used as the server deployment quantity of the shopping guide dialogue service;
[0024] According to the server deployment amounts of various shopping guide dialogue services, the server deployment amounts of each shopping guide dialogue service are normalized to obtain a service score of the shopping guide dialogue service;
[0025] According to the number of shopping guide dialogue services deployed by each server, the number of shopping guide dialogue services deployed by each server is normalized to obtain a server score of the server;
[0026] Use the following formula to calculate the user experience weight of each shopping guide dialogue service:
[0027]
[0028] in, The user experience weight of the i-th shopping guide dialogue service, is the number of servers deployed with the i-th shopping guide dialogue service, Serving the i-th shopping guide dialogue, For a server that deploys the i-th shopping guide dialogue service, is the service score of the i-th shopping guide dialogue service, is the server score of the j-th server where the i-th shopping guide dialogue service is deployed.
[0029] Optionally, determining a user experience score for each shopping guide dialogue service of the e-commerce service system includes:
[0030] For each shopping guide dialogue service, on a target server where the shopping guide dialogue service is deployed, a shopping guide dialogue flow text for guiding a user to shop is obtained, wherein the shopping guide dialogue flow text includes a plurality of question texts of the user based on shopping needs, and a reply text corresponding to each question text;
[0031] For each shopping guide dialogue service, the intelligent big model corresponding to the shopping guide dialogue service is used on the target server to score the shopping guide dialogue flow text according to the response quality of all reply texts in the shopping guide dialogue flow text to their corresponding question texts, so as to obtain the user experience score of the shopping guide dialogue service.
[0032] A downgrade migration device for a shopping guide dialogue service is applied to an e-commerce service system, wherein the e-commerce service system includes a plurality of servers, and at least one shopping guide dialogue service is deployed in each server. The device includes:
[0033] A migration result optimization calculation unit, used for, when there is a server overload in the e-commerce service system, taking the sum of weighted scores of user experience of various shopping guide dialogue services of the e-commerce service system as the optimization goal and taking the current resource limitation of each server as the optimization constraint condition, determining the optimization migration result of the shopping guide dialogue services deployed in each server;
[0034] An optimization migration unit, configured to migrate the shopping guide dialogue services deployed in each server based on the optimization migration result;
[0035] A user experience score determination unit, used to determine a user experience score for each shopping guide dialogue service of the e-commerce service system;
[0036] A user experience weight calculation unit, used to calculate the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers;
[0037] The user experience weighted score calculation unit is used to multiply the user experience score of each shopping guide dialogue service by the user experience weight of the shopping guide dialogue service to obtain the user experience weighted score of the shopping guide dialogue service.
[0038] Optionally, the optimization objective is:
[0039]
[0040] in, The weighted score of the user experience of the i-th shopping guide dialogue service;
[0041] The optimization constraints are:
[0042]
[0043] in, For each server under the current resource constraints, Serving the i-th shopping guide dialogue, is the GPU resource occupied by the i-th shopping guide dialogue service on a single server, is the number of servers deployed with the i-th shopping guide dialogue service.
[0044] Optionally, the user experience weight determination unit includes:
[0045] The principal component analysis calculation unit is used to calculate the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers through the principal component analysis method.
[0046] Optionally, the user experience weight calculation unit includes:
[0047] A service deployment quantity calculation unit, used to use the number of servers deployed with each shopping guide dialogue service as the server deployment quantity of the shopping guide dialogue service;
[0048] A service score calculation unit, configured to normalize the server deployment amount of each shopping guide dialogue service according to the server deployment amount of various shopping guide dialogue services, and obtain a service score of the shopping guide dialogue service;
[0049] A server score calculation unit, for normalizing the number of shopping guide dialogue services deployed by each server according to the number of shopping guide dialogue services deployed by each server, to obtain a server score for the server;
[0050] The weight calculation unit is used to calculate the user experience weight of each shopping guide dialogue service using the following formula:
[0051]
[0052] in, The user experience weight of the i-th shopping guide dialogue service, is the number of servers deployed with the i-th shopping guide dialogue service, Serving the i-th shopping guide dialogue, For a server that deploys the i-th shopping guide dialogue service, is the service score of the i-th shopping guide dialogue service, is the server score of the j-th server where the i-th shopping guide dialogue service is deployed.
[0053] Optionally, the user experience score calculation unit includes:
[0054] A dialogue flow text acquisition unit is used to acquire, for each shopping guide dialogue service, a shopping guide dialogue flow text guiding the user to shop on a target server where the shopping guide dialogue service is deployed, wherein the shopping guide dialogue flow text includes a plurality of question texts of the user based on shopping needs and a reply text corresponding to each question text;
[0055] The scoring unit is used to score the shopping guide dialogue flow text according to the quality of all reply texts in the shopping guide dialogue flow text to their corresponding question texts, using the intelligent big model corresponding to the shopping guide dialogue service on the target server for each shopping guide dialogue service, so as to obtain the user experience score of the shopping guide dialogue service.
[0056] A downgrade migration device for a shopping guide dialogue service, comprising a memory and a processor;
[0057] The memory is used to store programs;
[0058] The processor is used to execute the program to implement the various steps of the downgrade migration method of the shopping guide dialogue service as described above.
[0059] A storage medium stores a computer program, which, when executed by a processor, implements the various steps of the downgrade migration method of the shopping guide dialogue service as described above.
[0060] By means of the above technical solution, the present application obtains the user experience score of the system's shopping guide dialogue service, and uses the principal component analysis method to calculate the user experience weight of the shopping guide dialogue service according to the deployment of various shopping guide dialogue services in each server, so as to obtain the user experience weighted score. When the system has a server overload, the sum of the weighted scores of the user experience of each service is maximized as the optimization goal, and the shopping guide dialogue services deployed in each server are migrated based on the current resource limitations of each server. It can be seen that in the face of priority and limited resources and sudden increases in traffic, by optimizing the deployment of shopping guide dialogue services in each server, service migration is performed, so that various important shopping guide dialogue services can be smoothly provided to users under the optimization goal, so as to keep the user experience of shopping guide services from decreasing. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:
[0062] Figure 1 A schematic diagram of a process for implementing downgrade migration of a shopping guide dialogue service provided in an embodiment of the present application;
[0063] Figure 2 A schematic diagram of a flow chart for calculating a weighted score of a user experience of a shopping guide dialogue service provided in an embodiment of the present application;
[0064] Figure 3 A schematic diagram of the structure of a device for implementing downgrade migration of a shopping guide dialogue service provided in an embodiment of the present application;
[0065] Figure 4 A schematic diagram of the structure of a device for implementing downgrade migration of a shopping guide dialogue service provided in an embodiment of the present application. DETAILED DESCRIPTION
[0066] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0067] The solution of the present application can be implemented based on a terminal with data processing capabilities, and the terminal can be an e-commerce service system. The e-commerce service system can include multiple servers, and at least one shopping guide dialogue service is deployed in each server.
[0068] Specifically, the e-commerce service system can provide shopping guide dialogue services with various information, such as weather recommendations, thermal power products, etc. The collection of various shopping guide dialogue services can be: , represents the i-th shopping guide dialogue service, the i-th shopping guide dialogue service Deployed on superior, is the server set where the i-th shopping guide dialogue service is deployed. is the number of servers deployed with the i-th shopping guide dialogue service, Indicates the jth server where the i-th shopping guide dialogue service is deployed.
[0069] Next, combine Figure 1 The downgrade migration method of the shopping guide dialogue service of the present application may include the following steps:
[0070] Step S110: When there is a server overload in the e-commerce service system, the optimization goal is to maximize the sum of the weighted scores of user experience of various shopping guide dialogue services of the e-commerce service system, and the current resource limitations of each server are used as optimization constraints to determine the optimization migration results of the shopping guide dialogue services deployed in each server.
[0071] Specifically, the optimized migration result may be expressed as an optimal solution with maximizing the sum of weighted scores of user experience as the optimization goal and with the current resource limitations of each server as the constraint.
[0072] It is understandable that when there is a server overload in the e-commerce service system, service migration, scheduling or suspension is required according to the real-time resource constraints. Since the optimization goal is to maximize the sum of the weighted scores of the user experience, the optimized migration results can maintain the user experience of the shopping guide assistant service.
[0073] Step S120: Based on the optimization migration result, the shopping guide dialogue service deployed in each server is migrated.
[0074] Specifically, the optimization migration result may represent instruction information mainly based on transferring or suspending the shopping guide dialogue service deployed in the overloaded server.
[0075] It is understandable that when a server is overloaded, if other servers currently have sufficient resources, the shopping guide dialogue service deployed in the overloaded server can be transferred to other servers, or unimportant shopping guide dialogue services deployed in the overloaded server can be suspended. In the process of service migration, in order to better schedule the resources of each server, the shopping guide dialogue service deployed on the non-overloaded server can also be scheduled and migrated.
[0076] For example, the e-commerce service system has three servers A, B, and C. Server A is deployed with two shopping guide dialogue services (a and b), server B is deployed with one shopping guide dialogue service (a), and server C is deployed with five shopping guide dialogue services (a, b, c, y, and z, ranked in order of importance: a>b>c>y>z). During a big promotion, due to a sudden increase in traffic, server C is overloaded, server A is close to full load, and server B is sufficient. In this case, if server B can handle it, shopping guide dialogue services b and c deployed on server C can be migrated to server B. If servers B and C are also close to full load at this time, the least important shopping guide dialogue service z can be temporarily suspended.
[0077] The calculation process of the user experience weighted score of each shopping guide dialogue service of the e-commerce service system is as follows: Figure 2 As shown, the calculation process may include:
[0078] Step S210: Determine the user experience score of each shopping guide dialogue service of the e-commerce service system.
[0079] Specifically, the user experience score may indicate the user's preference for the shopping guide dialogue service. The higher the user's preference, the higher the user experience score, and vice versa.
[0080] Step S220: Calculate the user experience weight of each shopping guide dialogue service according to the deployment status of various shopping guide dialogue services in various servers.
[0081] Specifically, the user experience weight can represent the importance of the shopping guide dialogue service in the e-commerce service system, which can be analyzed through the principal component analysis method.
[0082] It is understandable that the more servers the shopping guide dialogue service is deployed on in the e-commerce service system, the more obvious the demand for the shopping guide dialogue service is, and therefore the more important the shopping guide dialogue service is. The more other shopping guide dialogue services are deployed on the server, the greater the traffic demand of the server is. The more important the server is, the more important the shopping guide dialogue service deployed on it is. The more important the shopping guide dialogue service is in the e-commerce service system, the greater the value of the user experience weight is, and vice versa.
[0083] Step S230: multiply the user experience score of each shopping guide dialogue service by the user experience weight of the shopping guide dialogue service to obtain a weighted user experience score of the shopping guide dialogue service.
[0084] The method for downgrading and migrating the shopping guide dialogue service provided in this embodiment obtains the user experience score of the shopping guide dialogue service of the system, and uses the principal component analysis method to calculate the user experience weight of the shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers to obtain the user experience weighted score. When the system has a server overload, the sum of the weighted scores of the user experience of each service is maximized as the optimization goal, and the shopping guide dialogue services deployed in each server are migrated with the current resource limit of each server as the constraint. It can be seen that in the face of priority and limited resources and sudden increase in traffic, by optimizing the deployment of the shopping guide dialogue service in each server, service migration is performed, so that various important shopping guide dialogue services can be smoothly provided to users under the optimization goal, so as to keep the user experience of the shopping guide service from being reduced.
[0085] In some embodiments of the present application, the optimization objectives and optimization constraints mentioned in the above embodiments are further introduced. Specifically, the optimization objectives may be:
[0086]
[0087] in, The weighted score of the user experience of the i-th shopping guide dialogue service.
[0088] The optimization constraints are:
[0089]
[0090] in, For each server under the current resource constraints, Serving the i-th shopping guide dialogue, is the GPU resource occupied by the i-th shopping guide dialogue service on a single server, is the number of servers deployed with the i-th shopping guide dialogue service.
[0091] It is understandable that when a server is overloaded in the e-commerce service system, various shopping guide dialogue services are all online services, occupying GPU resources on the servers on which they are deployed, and the same shopping guide dialogue service occupies the same GPU resources on different servers.
[0092] Specifically, the e-commerce service system can migrate and optimize the shopping guide dialogue services deployed in each server without exceeding the current resource limitations of each server. When the GPU resource utilization is maximized, that is, when the optimization constraints are met, the shopping guide dialogue services of each server can be fully utilized to maximize the user experience of the shopping guide service.
[0093] In some embodiments of the present application, the process of calculating the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in each server in step S220 is introduced, which may specifically include the following two methods:
[0094] The first method is to calculate the user experience weight of each shopping guide dialogue service based on the deployment of various shopping guide dialogue services in various servers through principal component analysis.
[0095] Specifically, the principal component analysis method may include a minimum projected distance method and a maximum projected variance method.
[0096] The second type, S1, uses the number of servers deployed with each shopping guide dialogue service as the server deployment quantity of the shopping guide dialogue service.
[0097] S2. According to the server deployment amounts of various shopping guide dialogue services, the server deployment amount of each shopping guide dialogue service is normalized to obtain a service score of the shopping guide dialogue service.
[0098] It is understandable that the more servers are deployed, the more obvious the demand for the shopping guide dialogue service is, and the more important the shopping guide dialogue service is. Therefore, the service score of this shopping guide dialogue service is higher. It can be seen that the service score represents the importance of the shopping guide dialogue service.
[0099] S3. According to the number of shopping guide dialogue services deployed by each server, the number of shopping guide dialogue services deployed by each server is normalized to obtain a server score of the server.
[0100] Specifically, the more types of shopping guide dialogue services deployed by the server, the greater the traffic demand of the server, and the more important the server is, and the higher the server score of the server is. It can be seen that the server score represents the importance of the server.
[0101] S4. Calculate the user experience weight of each shopping guide dialogue service using the following formula:
[0102]
[0103] in, The user experience weight of the i-th shopping guide dialogue service, is the number of servers deployed with the i-th shopping guide dialogue service, Serving the i-th shopping guide dialogue, For a server that deploys the i-th shopping guide dialogue service, is the service score of the i-th shopping guide dialogue service, is the server score of the j-th server where the i-th shopping guide dialogue service is deployed.
[0104] Specifically, the shopping guide dialogue services of each server where the shopping guide dialogue service is deployed are accumulated, and the obtained sum is multiplied by the service score of the shopping guide dialogue service to obtain the user experience weight of the shopping guide dialogue service, which can reflect the importance of the shopping guide dialogue service after comprehensive consideration of each server.
[0105] The downgrade migration method for the shopping guide dialogue service provided in this embodiment obtains the user experience weight of the shopping guide dialogue service through principal component analysis. It also comprehensively considers the distribution of the shopping guide dialogue service on the server and the distribution of the server on the shopping guide dialogue service, thereby obtaining the user experience weight of the shopping guide dialogue service, making the user experience weight result more reliable.
[0106] In some embodiments of the present application, the process of determining the user experience score of each shopping guide dialogue service of the e-commerce service system in step S210 is introduced, and the process may include:
[0107] S1. For each shopping guide dialogue service, on a target server where the shopping guide dialogue service is deployed, a shopping guide dialogue flow text for guiding a user to shop is obtained.
[0108] The shopping guide dialogue flow text may include multiple question texts based on shopping needs of users, and the corresponding answer texts for each question text. The answer texts are obtained by the intelligent big model called by the target server according to the question texts, and are displayed by the shopping guide terminal connected to the target server.
[0109] With QQ representing the user and AA representing the shopping guide terminal, an example of the shopping guide dialogue flow text is as follows:
[0110] QQ: I want to buy shoes
[0111] AA: No, I have clothes. Do you want some?
[0112] QQ: OK, what styles do you recommend?
[0113] AA: Dear, Chanel style, do you want to take a look? I bought some very good ones now, they are very suitable for you. I will help you take a look at those ones.
[0114] S2. For each shopping guide dialogue service, the intelligent big model corresponding to the shopping guide dialogue service is used on the target server to score the shopping guide dialogue flow text according to the quality of responses of all response texts in the shopping guide dialogue flow text to their corresponding question texts, so as to obtain the user experience score of the shopping guide dialogue service.
[0115] Among them, the intelligent big model can be an open source prompt engineering big model.
[0116] It is understandable that the user experience score is a direct reflection of the shopping guide's response. A high-scoring response text indicates that the response quality is high and can be used as a reference for the shopping guide's response. A low-scoring response text indicates that the response quality is low and can also be fed back to the intelligent big model for further optimization and as a sample for the intelligent big model's enhanced training.
[0117] Furthermore, the intelligent big model can be used to perform scoring on other target servers of the shopping guide dialogue service, and the scoring scores obtained by the respective target servers can be averaged as the final user experience score of the shopping guide dialogue service.
[0118] It is understandable that since different servers serve different user objects and have different question texts, different servers reply to different users with different texts. It can be seen that different servers obtain different user experience scores when evaluating the users they serve. In this case, all rating scores can be combined and the average value can be taken as the final user experience score, making the user experience score more reliable.
[0119] The following describes an apparatus for implementing downgrade migration of a shopping guide dialogue service provided in an embodiment of the present application. The apparatus for implementing downgrade migration of a shopping guide dialogue service described below and the method for implementing downgrade migration of a shopping guide dialogue service described above can be referenced to each other.
[0120] See also Figure 3 , Figure 3 A schematic diagram of the structure of a device for implementing downgrade migration of a shopping guide dialogue service disclosed in an embodiment of the present application.
[0121] like Figure 3 As shown, the device may include:
[0122] The migration result optimization calculation unit 11 is used to, when there is a server overload in the e-commerce service system, take the sum of the weighted scores of user experience of various shopping guide dialogue services of the e-commerce service system as the optimization goal, take the current resource limitation of each server as the optimization constraint condition, and determine the optimization migration result of the shopping guide dialogue service deployed in each server;
[0123] An optimization migration unit 12, configured to migrate the shopping guide dialogue services deployed in each server based on the optimization migration result;
[0124] A user experience score determination unit 13, configured to determine a user experience score for each shopping guide dialogue service of the e-commerce service system;
[0125] The user experience weight calculation unit 14 is used to calculate the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers;
[0126] The user experience weighted score calculation unit 15 is used to multiply the user experience score of each shopping guide dialogue service by the user experience weight of the shopping guide dialogue service to obtain the user experience weighted score of the shopping guide dialogue service.
[0127] Optionally, the optimization objective is:
[0128]
[0129] in, The weighted score of the user experience of the i-th shopping guide dialogue service;
[0130] The optimization constraints are:
[0131]
[0132] in, For each server under the current resource constraints, Serving the i-th shopping guide dialogue, is the GPU resource occupied by the i-th shopping guide dialogue service on a single server, is the number of servers deployed with the i-th shopping guide dialogue service.
[0133] Optionally, the user experience weight determination unit includes:
[0134] The principal component analysis calculation unit is used to calculate the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers through the principal component analysis method.
[0135] Optionally, the user experience weight calculation unit includes:
[0136] A service deployment quantity calculation unit, used to use the number of servers deployed with each shopping guide dialogue service as the server deployment quantity of the shopping guide dialogue service;
[0137] A service score calculation unit, configured to normalize the server deployment amount of each shopping guide dialogue service according to the server deployment amount of various shopping guide dialogue services, and obtain a service score of the shopping guide dialogue service;
[0138] A server score calculation unit, for normalizing the number of shopping guide dialogue services deployed by each server according to the number of shopping guide dialogue services deployed by each server, to obtain a server score for the server;
[0139] The weight calculation unit is used to calculate the user experience weight of each shopping guide dialogue service using the following formula:
[0140]
[0141] in, The user experience weight of the i-th shopping guide dialogue service, is the number of servers deployed with the i-th shopping guide dialogue service, Serving the i-th shopping guide dialogue, For a server that deploys the i-th shopping guide dialogue service, is the service score of the i-th shopping guide dialogue service, is the server score of the j-th server where the i-th shopping guide dialogue service is deployed.
[0142] Optionally, the user experience score calculation unit includes:
[0143] A dialogue flow text acquisition unit is used to acquire, for each shopping guide dialogue service, a shopping guide dialogue flow text guiding the user to shop on a target server where the shopping guide dialogue service is deployed, wherein the shopping guide dialogue flow text includes a plurality of question texts of the user based on shopping needs and a reply text corresponding to each question text;
[0144] The scoring unit is used to score the shopping guide dialogue flow text according to the quality of all reply texts in the shopping guide dialogue flow text to their corresponding question texts, using the intelligent big model corresponding to the shopping guide dialogue service on the target server for each shopping guide dialogue service, so as to obtain the user experience score of the shopping guide dialogue service.
[0145] The apparatus for downgrading and migrating the shopping guide dialogue service provided in the embodiment of the present application can be applied to a device for downgrading and migrating the shopping guide dialogue service, such as an e-commerce service system. Figure 4 The hardware structure diagram of the device for downgrade migration of the shopping guide dialogue service is shown, referring to Figure 4 , the hardware structure of the device for downgrade migration of the shopping guide dialogue service may include: at least one processor 1, at least one communication interface 2, at least one memory 3 and at least one communication bus 4;
[0146] In the embodiment of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 communicate with each other through the communication bus 4;
[0147] The processor 1 may be a central processing unit CPU, or an application-specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention, etc.;
[0148] The memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), etc., such as at least one disk memory;
[0149] The memory stores a program, and the processor can call the program stored in the memory, wherein the program is used to:
[0150] When there is a server overload in the e-commerce service system, the optimization goal is to maximize the sum of the weighted scores of user experience of various shopping guide dialogue services of the e-commerce service system, and the current resource limit of each server is used as the optimization constraint condition to determine the optimization migration result of the shopping guide dialogue service deployed in each server;
[0151] Based on the optimization migration result, the shopping guide dialogue service deployed in each server is migrated;
[0152] The calculation process of the user experience weighted score of each shopping guide dialogue service of the e-commerce service system is as follows:
[0153] Determine a user experience score for each shopping guide dialogue service of the e-commerce service system;
[0154] Calculate the user experience weight of each shopping guide dialogue service based on the deployment of various shopping guide dialogue services in various servers;
[0155] The user experience score of each shopping guide dialogue service is multiplied by the user experience weight of the shopping guide dialogue service to obtain a weighted user experience score of the shopping guide dialogue service.
[0156] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0157] The embodiment of the present application further provides a storage medium, which may store a program suitable for execution by a processor, wherein the program is used to:
[0158] When there is a server overload in the e-commerce service system, the optimization goal is to maximize the sum of the weighted scores of user experience of various shopping guide dialogue services of the e-commerce service system, and the current resource limit of each server is used as the optimization constraint condition to determine the optimization migration result of the shopping guide dialogue service deployed in each server;
[0159] Based on the optimization migration result, the shopping guide dialogue service deployed in each server is migrated;
[0160] The calculation process of the user experience weighted score of each shopping guide dialogue service of the e-commerce service system is as follows:
[0161] Determine a user experience score for each shopping guide dialogue service of the e-commerce service system;
[0162] Calculate the user experience weight of each shopping guide dialogue service based on the deployment of various shopping guide dialogue services in various servers;
[0163] The user experience score of each shopping guide dialogue service is multiplied by the user experience weight of the shopping guide dialogue service to obtain a weighted user experience score of the shopping guide dialogue service.
[0164] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0165] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0166] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can refer to each other.
[0167] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for downgrading and migrating a shopping guide dialogue service, characterized in that: Applied to an e-commerce service system, the e-commerce service system includes multiple servers, each of which is deployed with at least one shopping guide dialogue service, the method includes: When there is a server overload in the e-commerce service system, the optimization goal is to maximize the sum of the weighted scores of user experience of various shopping guide dialogue services of the e-commerce service system, and the current resource limit of each server is used as the optimization constraint condition to determine the optimization migration result of the shopping guide dialogue service deployed in each server; Based on the optimization migration result, the shopping guide dialogue service deployed in each server is migrated; The calculation process of the user experience weighted score of each shopping guide dialogue service of the e-commerce service system is as follows: Determine a user experience score for each shopping guide dialogue service of the e-commerce service system; Calculate the user experience weight of each shopping guide dialogue service based on the deployment of various shopping guide dialogue services in various servers; The user experience score of each shopping guide dialogue service is multiplied by the user experience weight of the shopping guide dialogue service to obtain a weighted user experience score of the shopping guide dialogue service.
2. The method according to claim 1, characterized in that The optimization goal is: in, The weighted score of the user experience of the i-th shopping guide dialogue service; The optimization constraints are: in, For each server under the current resource constraints, Serving the i-th shopping guide dialogue, is the GPU resource occupied by the i-th shopping guide dialogue service on a single server, is the number of servers deployed with the i-th shopping guide dialogue service.
3. The method according to claim 1, characterized in that The calculation of the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers includes: Through principal component analysis, the user experience weight of each shopping guide dialogue service is calculated according to the deployment of various shopping guide dialogue services in various servers.
4. The method according to claim 1, characterized in that: The calculation of the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers includes: The number of servers deployed with each shopping guide dialogue service is used as the server deployment quantity of the shopping guide dialogue service; According to the server deployment amounts of various shopping guide dialogue services, the server deployment amounts of each shopping guide dialogue service are normalized to obtain a service score of the shopping guide dialogue service; According to the number of shopping guide dialogue services deployed by each server, the number of shopping guide dialogue services deployed by each server is normalized to obtain a server score of the server; Use the following formula to calculate the user experience weight of each shopping guide dialogue service: in, The user experience weight of the i-th shopping guide dialogue service, is the number of servers deployed with the i-th shopping guide dialogue service, Serving the i-th shopping guide dialogue, For a server that deploys the i-th shopping guide dialogue service, is the service score of the i-th shopping guide dialogue service, is the server score of the j-th server where the i-th shopping guide dialogue service is deployed.
5. The method according to any one of claims 1 to 4, characterized in that: Determine the user experience score of each shopping guide dialogue service of the e-commerce service system, including: For each shopping guide dialogue service, on a target server where the shopping guide dialogue service is deployed, a shopping guide dialogue flow text for guiding a user to shop is obtained, wherein the shopping guide dialogue flow text includes a plurality of question texts of the user based on shopping needs, and a reply text corresponding to each question text; For each shopping guide dialogue service, the intelligent big model corresponding to the shopping guide dialogue service is used on the target server to score the shopping guide dialogue flow text according to the response quality of all reply texts in the shopping guide dialogue flow text to their corresponding question texts, so as to obtain the user experience score of the shopping guide dialogue service.
6. A downgrade migration device for a shopping guide dialogue service, characterized in that: Applied to an e-commerce service system, the e-commerce service system includes multiple servers, each of which is deployed with at least one shopping guide dialogue service, and the device includes: A migration result optimization calculation unit, used for, when there is a server overload in the e-commerce service system, taking the sum of weighted scores of user experience of various shopping guide dialogue services of the e-commerce service system as the optimization goal and taking the current resource limitation of each server as the optimization constraint condition, determining the optimization migration result of the shopping guide dialogue services deployed in each server; An optimization migration unit, configured to migrate the shopping guide dialogue services deployed in each server based on the optimization migration result; A user experience score calculation unit, used to determine a user experience score for each shopping guide dialogue service of the e-commerce service system; A user experience weight calculation unit, used to calculate the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers; The user experience weighted score calculation unit is used to multiply the user experience score of each shopping guide dialogue service by the user experience weight of the shopping guide dialogue service to obtain the user experience weighted score of the shopping guide dialogue service.
7. The device according to claim 6, characterized in that The optimization goal is: in, The weighted score of the user experience of the i-th shopping guide dialogue service; The optimization constraints are: in, For each server under the current resource constraints, Serving the i-th shopping guide dialogue, is the GPU resource occupied by the i-th shopping guide dialogue service on a single server, is the number of servers deployed with the i-th shopping guide dialogue service.
8. The device according to claim 6, characterized in that The user experience weight calculation unit includes: The principal component analysis calculation unit is used to calculate the user experience weight of each shopping guide dialogue service according to the deployment of various shopping guide dialogue services in various servers through the principal component analysis method.
9. A downgrade migration device for a shopping guide dialogue service, characterized in that: including memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the downgrade migration method of the shopping guide dialogue service as described in any one of claims 1 to 5.
10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the downgrade migration method of the shopping guide dialogue service as described in any one of claims 1 to 5 is implemented.
