Method and system for determining computing power scheduling strategy based on heterogeneous planning

Abstract

The present invention discloses a method and system for determining a computing power scheduling strategy based on heterogeneous planning, the method comprising: obtaining a computing power request set received by multiple network modules in the same computing power network; determining the request pressure corresponding to each network module according to the request parameters of all computing power requests of each computing power request set; determining the computing tolerance corresponding to each network module according to the historical computing records and module performance parameters corresponding to each network module; determining the computing power scheduling strategy corresponding to all the network modules according to the request pressure and the computing tolerance. It can be seen that the present invention can improve the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for determining a computing power scheduling strategy based on heterogeneous planning. Background Art

[0002] With the improvement and popularization of algorithm technologies such as neural networks, more and more users have begun to use online computing resources to perform customized algorithm training tasks or algorithm execution tasks. Correspondingly, more and more platforms have begun to integrate the computing resources of multiple different network modules online to provide users with corresponding computing services. This has put forward higher requirements on how computing platforms can effectively schedule the computing resources of different network modules to better match user computing needs. When performing computing power scheduling, most of the existing technologies only determine computing resources through direct user designation, or implement computing power scheduling through simple priority scheduling rules, and do not fully combine the historical operation records and performance parameters of the network module to schedule computing resources. Therefore, the scheduling effect is poor, the resource utilization rate is low, and it is impossible to provide users with a better computing power usage experience. It can be seen that the existing technology has defects that need to be solved urgently. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method and system for determining a computing power scheduling strategy based on heterogeneous planning, which can improve the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0004] In order to solve the above technical problems, the first aspect of the present invention discloses a method for determining a computing power scheduling strategy based on heterogeneous planning, the method comprising:

[0005] Get the computing power request set received by multiple network modules in the same computing power network;

[0006] Determine the request pressure corresponding to each of the network modules according to the request parameters of all the computing power requests of each of the computing power request sets;

[0007] Determine the operation tolerance corresponding to each network module according to the historical operation records and module performance parameters corresponding to each network module;

[0008] Determine the computing power scheduling strategies corresponding to all the network modules according to the request pressure and the computing tolerance.

[0009] As an optional embodiment, in the first aspect of the present invention, the request parameters include at least one of the requested task data volume, the requested task computing volume, the requested task type, the requested task data type, the requested purpose, the requested task application environment and the requested task programming language.

[0010] As an optional implementation, in the first aspect of the present invention, the historical operation records include a plurality of historical data transmission records between the network modules and a plurality of historical data calculation records of the network modules.

[0011] As an optional embodiment, in the first aspect of the present invention, the module performance parameters include hardware device parameters, hardware device models, connection methods between hardware devices, hardware device working collaboration methods, and historical hardware device occupancy, historical hardware device temperature and historical hardware device humidity corresponding to the network module.

[0012] As an optional implementation, in the first aspect of the present invention, determining the request pressure corresponding to each of the network modules according to the request parameters of all computing power requests of each of the computing power request sets includes:

[0013] For each computing power request of the computing power request set of each network module, the request parameters of the computing power request are input into a trained load prediction neural network to obtain a request load corresponding to the computing power request; the load prediction neural network is trained by a training data set including a plurality of training computing power request parameters and corresponding load annotations;

[0014] Calculate the weighted average of the request loads of all the computing power requests in the computing power request set to obtain the request pressure corresponding to the network module.

[0015] As an optional implementation, in the first aspect of the present invention, determining the operation tolerance corresponding to each network module according to the historical operation records and module performance parameters corresponding to each network module includes:

[0016] For each of the network modules, calculating a historical computing capability parameter of the network module according to the historical computing record corresponding to the network module;

[0017] Inputting the module performance parameters of the network module into a trained computing capacity prediction neural network to obtain the predicted computing capacity parameters corresponding to the network module; the computing capacity prediction neural network is trained by a training data set including a plurality of training module performance parameters and corresponding computing capacity annotations;

[0018] The weighted average value of the historical computing capability parameter and the predicted computing capability parameter is calculated to obtain the computing tolerance corresponding to the network module.

[0019] As an optional implementation, in the first aspect of the present invention, the calculating the historical computing capability parameter of the network module according to the historical computing record corresponding to the network module includes:

[0020] Calculate the average value of the transmission speed of each historical data transmission record in the historical operation record corresponding to the network module to obtain the data transmission parameter corresponding to the network module;

[0021] Calculate the average completion time of all data calculation tasks in the historical data calculation record corresponding to the network module to obtain the data calculation parameters corresponding to the network module;

[0022] The weighted average value of the data transmission parameter and the data calculation parameter is calculated to obtain the historical computing capacity parameter of the network module.

[0023] As an optional implementation, in the first aspect of the present invention, determining the computing power scheduling strategy corresponding to all the network modules according to the request pressure and the computing tolerance includes:

[0024] Calculating the ratio of the operation tolerance of each network module to the request pressure to obtain the corresponding load parameters of each network module;

[0025] The objective function is set to minimize the number of computing power requests allocated to each network module in the allocation scheme;

[0026] The setting restriction conditions include that the network module with a higher bearing parameter has a higher number of computing power requests allocated, the number of computing power requests allocated to each network module is less than a preset number threshold, and the network module with a bearing parameter lower than a preset parameter threshold does not allocate computing power requests;

[0027] Based on the dynamic programming algorithm, all computing power requests in the computing power request set are iteratively allocated to each of the network modules according to the objective function and the constraint conditions until the best allocation scheme is obtained to obtain the computing power scheduling strategy.

[0028] A second aspect of an embodiment of the present invention discloses a computing power scheduling strategy determination system based on heterogeneous planning, the system comprising:

[0029] An acquisition module is used to obtain a set of computing power requests received by multiple network modules in the same computing power network;

[0030] A first determination module, configured to determine the request pressure corresponding to each of the network modules according to the request parameters of all computing power requests of each computing power request set;

[0031] A second determination module, used to determine the operation tolerance corresponding to each of the network modules according to the historical operation records and module performance parameters corresponding to each of the network modules;

[0032] The third determination module is used to determine the computing power scheduling strategy corresponding to all the network modules according to the request pressure and the computing tolerance.

[0033] As an optional embodiment, in the second aspect of the present invention, the request parameters include at least one of the requested task data volume, the requested task computing volume, the requested task type, the requested task data type, the requested purpose, the requested task application environment and the requested task programming language.

[0034] As an optional implementation, in the second aspect of the present invention, the historical operation records include a plurality of historical data transmission records between the network modules and a plurality of historical data calculation records of the network modules.

[0035] As an optional embodiment, in the second aspect of the present invention, the module performance parameters include hardware device parameters, hardware device models, connection methods between hardware devices, hardware device working collaboration methods, and historical hardware device occupancy, historical hardware device temperature and historical hardware device humidity corresponding to the network module.

[0036] As an optional implementation, in the second aspect of the present invention, the first determination module determines the specific manner of the request pressure corresponding to each of the network modules according to the request parameters of all computing power requests of each computing power request set, including:

[0037] For each computing power request of the computing power request set of each network module, the request parameters of the computing power request are input into a trained load prediction neural network to obtain a request load corresponding to the computing power request; the load prediction neural network is trained by a training data set including a plurality of training computing power request parameters and corresponding load annotations;

[0038] Calculate the weighted average of the request loads of all the computing power requests in the computing power request set to obtain the request pressure corresponding to the network module.

[0039] As an optional implementation, in the second aspect of the present invention, the second determination module determines the specific manner of the operation tolerance corresponding to each of the network modules according to the historical operation records and module performance parameters corresponding to each of the network modules, including:

[0040] For each of the network modules, calculating a historical computing capability parameter of the network module according to the historical computing record corresponding to the network module;

[0041] Inputting the module performance parameters of the network module into a trained computing capacity prediction neural network to obtain the predicted computing capacity parameters corresponding to the network module; the computing capacity prediction neural network is trained by a training data set including a plurality of training module performance parameters and corresponding computing capacity annotations;

[0042] The weighted average value of the historical computing capability parameter and the predicted computing capability parameter is calculated to obtain the computing tolerance corresponding to the network module.

[0043] As an optional implementation, in the second aspect of the present invention, the specific manner in which the second determination module calculates the historical computing capability parameter of the network module according to the historical computing record corresponding to the network module includes:

[0044] Calculate the average value of the transmission speed of each historical data transmission record in the historical operation record corresponding to the network module to obtain the data transmission parameter corresponding to the network module;

[0045] Calculate the average completion time of all data calculation tasks in the historical data calculation record corresponding to the network module to obtain the data calculation parameters corresponding to the network module;

[0046] The weighted average value of the data transmission parameter and the data calculation parameter is calculated to obtain the historical computing capacity parameter of the network module.

[0047] As an optional implementation, in the second aspect of the present invention, the third determination module determines the specific manner of the computing power scheduling strategy corresponding to all the network modules according to the request pressure and the computing tolerance, including:

[0048] Calculating the ratio of the operation tolerance of each network module to the request pressure to obtain the corresponding load parameters of each network module;

[0049] The objective function is set to minimize the number of computing power requests allocated to each network module in the allocation scheme;

[0050] The setting restriction conditions include that the network module with a higher bearing parameter has a higher number of computing power requests allocated, the number of computing power requests allocated to each network module is less than a preset number threshold, and the network module with a bearing parameter lower than a preset parameter threshold does not allocate computing power requests;

[0051] Based on the dynamic programming algorithm, all computing power requests in the computing power request set are iteratively allocated to each of the network modules according to the objective function and the constraint conditions until the best allocation scheme is obtained to obtain the computing power scheduling strategy.

[0052] The third aspect of the present invention discloses another computing power scheduling strategy determination system based on heterogeneous planning, the system comprising:

[0053] A memory storing executable program code;

[0054] a processor coupled to the memory;

[0055] The processor calls the executable program code stored in the memory to execute part or all of the steps in the method for determining a computing power scheduling strategy based on heterogeneous planning disclosed in the first aspect of the present invention.

[0056] The fourth aspect of the present invention discloses a computer storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute part or all of the steps in the method for determining a computing power scheduling strategy based on heterogeneous planning disclosed in the first aspect of the present invention.

[0057] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0058] The present invention can determine the corresponding request pressure based on the computing power request set received by the network module, and then determine the computing tolerance based on the historical computing records and performance parameters of the network module, so as to comprehensively determine a more reasonable and stable computing power scheduling strategy, thereby improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improving the utilization rate of computing power resources, and providing users with a better computing power usage experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0060] Figure 1 It is a flow chart of a method for determining a computing power scheduling strategy based on heterogeneous planning disclosed in an embodiment of the present invention.

[0061] Figure 2 It is a structural diagram of a computing power scheduling strategy determination system based on heterogeneous planning disclosed in an embodiment of the present invention.

[0062] Figure 3 It is a structural diagram of another system for determining computing power scheduling strategies based on heterogeneous planning disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or equipment.

[0065] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0066] The present invention discloses a method and system for determining a computing power scheduling strategy based on heterogeneous planning, which can determine the corresponding request pressure based on the computing power request set received by the network module, and then determine the computing tolerance based on the historical computing records and performance parameters of the network module, so as to comprehensively determine a more reasonable and stable computing power scheduling strategy, thereby improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improving the utilization rate of computing power resources, and providing users with a better computing power usage experience. The following are detailed descriptions.

[0067] Embodiment 1

[0068] See also Figure 1 , Figure 1 : is a flow chart of a method for determining a computing power scheduling strategy based on heterogeneous planning disclosed in an embodiment of the present invention. Figure 1 The described method for determining computing power scheduling strategy based on heterogeneous planning can be applied to a data processing system / data processing device / data processing server (wherein the server includes a local processing server or a cloud processing server). Figure 1 As shown, the method for determining a computing power scheduling strategy based on heterogeneous planning may include the following operations:

[0069] 101. Obtain a set of computing power requests received by multiple network modules in the same computing power network.

[0070] 102. Determine the request pressure corresponding to each network module according to the request parameters of all computing power requests of each computing power request set.

[0071] 103. Determine the operation tolerance corresponding to each network module according to the historical operation records and module performance parameters corresponding to each network module.

[0072] 104. Determine the computing power scheduling strategy corresponding to all network modules based on the request pressure and computing tolerance.

[0073] It can be seen that the above-mentioned embodiments of the invention can determine the corresponding request pressure based on the computing power request set received by the network module, and then determine the computing tolerance based on the historical computing records and performance parameters of the network module, so as to comprehensively determine a more reasonable and stable computing power scheduling strategy, thereby improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improving the utilization rate of computing power resources, and providing users with a better computing power usage experience.

[0074] As an optional embodiment, in the above steps, the request parameters include at least one of the requested task data volume, requested task computing volume, requested task type, requested task data type, requested purpose, requested task application environment and requested task programming language.

[0075] It can be seen that through the above optional embodiments, the content of the request parameters is clarified, which can more accurately characterize the characteristics and requirements of the user's computing power request, facilitate subsequent computing power scheduling, and assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0076] As an optional embodiment, in the above steps, the historical operation records include multiple historical data transmission records between network modules and multiple historical data calculation records of the network modules.

[0077] It can be seen that through the above optional embodiments, the content of the historical calculation records is clarified, which can more accurately characterize the historical data transmission rules and historical data calculation rules of the network module, so as to facilitate the subsequent computing power scheduling, assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0078] As an optional embodiment, in the above steps, the module performance parameters include hardware device parameters within the network module, hardware device models, connection methods between hardware devices, hardware device working collaboration methods, and historical hardware device occupancy, historical hardware device temperature, and historical hardware device humidity corresponding to the network module.

[0079] It can be seen that through the above optional embodiments, the content of the module performance parameters is limited, which can more accurately characterize the characteristics and information related to the performance of the network module, so as to facilitate the subsequent computing power scheduling, assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0080] As an optional embodiment, in the above steps, determining the request pressure corresponding to each network module according to the request parameters of all computing power requests of each computing power request set includes:

[0081] For each computing power request of the computing power request set of each network module, the request parameters of the computing power request are input into the trained load prediction neural network to obtain the request load corresponding to the computing power request; optionally, the load prediction neural network is trained by a training data set including a plurality of training computing power request parameters and corresponding load annotations;

[0082] Calculate the weighted average of the request loads of all computing power requests in the computing power request set to obtain the request pressure corresponding to the network module.

[0083] It can be seen that through the above optional embodiments, it is possible to predict the load corresponding to each computing power request based on the load prediction neural network, and then obtain the request pressure corresponding to each network module based on weighted calculation, so as to facilitate subsequent computing power scheduling, and assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0084] As an optional embodiment, in the above steps, determining the operation tolerance corresponding to each network module according to the historical operation records and module performance parameters corresponding to each network module includes:

[0085] For each network module, calculate the historical computing capacity parameter of the network module according to the historical computing record corresponding to the network module;

[0086] Inputting the module performance parameters of the network module into a trained computing power prediction neural network to obtain the predicted computing power parameters corresponding to the network module; optionally, the computing power prediction neural network is trained by a training data set including a plurality of training module performance parameters and corresponding computing power annotations;

[0087] The weighted average value of the historical computing power parameters and the predicted computing power parameters is calculated to obtain the computing tolerance corresponding to the network module.

[0088] It can be seen that through the above-mentioned optional embodiments, it is possible to calculate the historical computing power parameters according to the historical computing records of the network module, and after predicting the computing power parameters based on the module performance parameters based on the neural network, comprehensively calculate the computing tolerance corresponding to the network module, so as to facilitate subsequent computing power scheduling, and assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0089] As an optional embodiment, in the above step, calculating the historical computing capability parameter of the network module according to the historical computing record corresponding to the network module includes:

[0090] Calculate the average value of the transmission speed of each historical data transmission record in the historical operation record corresponding to the network module to obtain the data transmission parameter corresponding to the network module;

[0091] Calculate the average completion time of all data calculation tasks in the historical data calculation record corresponding to the network module to obtain the data calculation parameters corresponding to the network module;

[0092] The weighted average value of the data transmission parameter and the data calculation parameter is calculated to obtain the historical computing capacity parameter of the network module.

[0093] It can be seen that through the above-mentioned optional embodiments, the historical computing capacity parameters can be accurately calculated according to the parameters in the historical data transmission records and data calculation records in the historical computing records of the network module, so as to facilitate the subsequent comprehensive calculation of the computing tolerance corresponding to the network module, and assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0094] As an optional embodiment, in the above steps, determining the computing power scheduling strategy corresponding to all network modules according to the request pressure and computing tolerance includes:

[0095] Calculate the ratio of the operation tolerance and request pressure of each network module to obtain the corresponding load parameters of each network module;

[0096] The objective function is set to minimize the number of computing power requests allocated to each network module in the allocation scheme;

[0097] The setting restriction conditions include that a network module with a higher carrying parameter is allocated a higher number of computing power requests, the number of computing power requests allocated to each network module is less than a preset number threshold, and a network module with a carrying parameter lower than a preset parameter threshold is not allocated computing power requests;

[0098] Based on the dynamic programming algorithm, the computing power requests in all computing power request sets are iteratively allocated to each network module according to the objective function and constraints until the best allocation plan is obtained to obtain the computing power scheduling strategy.

[0099] It can be seen that through the above optional embodiments, computing power requests can be reasonably allocated according to the carrying parameters of the network modules and the dynamic programming algorithm, so as to improve the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0100] Embodiment 2

[0101] See also Figure 2 , Figure 2 : is a schematic diagram of a system for determining a computing power scheduling strategy based on heterogeneous planning disclosed in an embodiment of the present invention. Figure 2 The computing power scheduling strategy determination system based on heterogeneous planning described above can be applied to a data processing system / data processing device / data processing server (wherein the server includes a local processing server or a cloud processing server). Figure 2 As shown, the computing power scheduling strategy determination system based on heterogeneous planning may include:

[0102] The acquisition module 201 is used to obtain a set of computing power requests received by multiple network modules in the same computing power network.

[0103] The first determination module 202 is used to determine the request pressure corresponding to each network module according to the request parameters of all computing power requests of each computing power request set.

[0104] The second determination module 203 is used to determine the operation tolerance corresponding to each network module according to the historical operation records and module performance parameters corresponding to each network module.

[0105] The third determination module 204 is used to determine the computing power scheduling strategy corresponding to all network modules according to the request pressure and computing tolerance.

[0106] It can be seen that the above-mentioned embodiments of the invention can determine the corresponding request pressure based on the computing power request set received by the network module, and then determine the computing tolerance based on the historical computing records and performance parameters of the network module, so as to comprehensively determine a more reasonable and stable computing power scheduling strategy, thereby improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improving the utilization rate of computing power resources, and providing users with a better computing power usage experience.

[0107] As an optional embodiment, the request parameters include at least one of the requested task data volume, the requested task computing volume, the requested task type, the requested task data type, the requested purpose, the requested task application environment and the requested task programming language.

[0108] It can be seen that through the above optional embodiments, the content of the request parameters is clarified, which can more accurately characterize the characteristics and requirements of the user's computing power request, facilitate subsequent computing power scheduling, and assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0109] As an optional embodiment, the historical operation records include multiple historical data transmission records between network modules and multiple historical data calculation records of the network modules.

[0110] It can be seen that through the above optional embodiments, the content of the historical calculation records is clarified, which can more accurately characterize the historical data transmission rules and historical data calculation rules of the network module, so as to facilitate the subsequent computing power scheduling, assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0111] As an optional embodiment, the module performance parameters include hardware device parameters within the network module, hardware device models, connection methods between hardware devices, hardware device working coordination methods, and historical hardware device occupancy, historical hardware device temperature, and historical hardware device humidity corresponding to the network module.

[0112] It can be seen that through the above optional embodiments, the content of the module performance parameters is limited, which can more accurately characterize the characteristics and information related to the performance of the network module, so as to facilitate the subsequent computing power scheduling, assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0113] As an optional embodiment, the first determination module determines the specific manner of the request pressure corresponding to each network module according to the request parameters of all computing power requests of each computing power request set, including:

[0114] For each computing power request of the computing power request set of each network module, the request parameters of the computing power request are input into the trained load prediction neural network to obtain the request load corresponding to the computing power request; optionally, the load prediction neural network is trained by a training data set including a plurality of training computing power request parameters and corresponding load annotations;

[0115] Calculate the weighted average of the request loads of all computing power requests in the computing power request set to obtain the request pressure corresponding to the network module.

[0116] It can be seen that through the above optional embodiments, it is possible to predict the load corresponding to each computing power request based on the load prediction neural network, and then obtain the request pressure corresponding to each network module based on weighted calculation, so as to facilitate subsequent computing power scheduling, and assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0117] As an optional embodiment, the second determination module determines the specific manner of the operation tolerance corresponding to each network module according to the historical operation records and module performance parameters corresponding to each network module, including:

[0118] For each network module, calculate the historical computing capacity parameter of the network module according to the historical computing record corresponding to the network module;

[0119] Inputting the module performance parameters of the network module into a trained computing power prediction neural network to obtain the predicted computing power parameters corresponding to the network module; optionally, the computing power prediction neural network is trained by a training data set including a plurality of training module performance parameters and corresponding computing power annotations;

[0120] The weighted average value of the historical computing power parameters and the predicted computing power parameters is calculated to obtain the computing tolerance corresponding to the network module.

[0121] It can be seen that through the above-mentioned optional embodiments, it is possible to calculate the historical computing power parameters according to the historical computing records of the network module, and after predicting the computing power parameters based on the module performance parameters based on the neural network, comprehensively calculate the computing tolerance corresponding to the network module, so as to facilitate subsequent computing power scheduling, and assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0122] As an optional embodiment, the specific manner in which the second determination module calculates the historical computing capability parameter of the network module according to the historical computing record corresponding to the network module includes:

[0123] Calculate the average value of the transmission speed of each historical data transmission record in the historical operation record corresponding to the network module to obtain the data transmission parameter corresponding to the network module;

[0124] Calculate the average completion time of all data calculation tasks in the historical data calculation record corresponding to the network module to obtain the data calculation parameters corresponding to the network module;

[0125] The weighted average value of the data transmission parameter and the data calculation parameter is calculated to obtain the historical computing capacity parameter of the network module.

[0126] It can be seen that through the above-mentioned optional embodiments, the historical computing capacity parameters can be accurately calculated according to the parameters in the historical data transmission records and data calculation records in the historical computing records of the network module, so as to facilitate the subsequent comprehensive calculation of the computing tolerance corresponding to the network module, and assist in improving the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0127] As an optional embodiment, the third determination module determines the specific manner of the computing power scheduling strategy corresponding to all network modules according to the request pressure and the computing tolerance, including:

[0128] Calculate the ratio of the operation tolerance and request pressure of each network module to obtain the corresponding load parameters of each network module;

[0129] The objective function is set to minimize the number of computing power requests allocated to each network module in the allocation scheme;

[0130] The setting restriction conditions include that a network module with a higher carrying parameter is allocated a higher number of computing power requests, the number of computing power requests allocated to each network module is less than a preset number threshold, and a network module with a carrying parameter lower than a preset parameter threshold is not allocated computing power requests;

[0131] Based on the dynamic programming algorithm, the computing power requests in all computing power request sets are iteratively allocated to each network module according to the objective function and constraints until the best allocation plan is obtained to obtain the computing power scheduling strategy.

[0132] It can be seen that through the above optional embodiments, computing power requests can be reasonably allocated according to the carrying parameters of the network modules and the dynamic programming algorithm, so as to improve the computing power scheduling effect and efficiency of different network modules in the computing power network, improve the utilization rate of computing power resources, and provide users with a better computing power usage experience.

[0133] Embodiment 3

[0134] See also Figure 3 , Figure 3It is another computing power scheduling strategy determination system based on heterogeneous planning disclosed in an embodiment of the present invention. Figure 3 The computing power scheduling strategy determination system based on heterogeneous planning is applied to a data processing system / data processing device / data processing server (wherein the server includes a local processing server or a cloud processing server). Figure 3 As shown, the computing power scheduling strategy determination system based on heterogeneous planning may include:

[0135] A memory 301 storing executable program codes;

[0136] a processor 302 coupled to the memory 301;

[0137] The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the method for determining a computing power scheduling strategy based on heterogeneous planning described in the first embodiment.

[0138] Embodiment 4

[0139] An embodiment of the present invention discloses a computer-readable storage medium that stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the method for determining a computing power scheduling strategy based on heterogeneous planning described in Embodiment 1.

[0140] Embodiment 5

[0141] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps of the method for determining a computing power scheduling strategy based on heterogeneous planning described in Example 1.

[0142] The above describes specific embodiments of the present specification, and other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily have to be performed in the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0143] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0144] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0145] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification may be in the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the embodiments of this specification may be in 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 contain computer-usable program code.

[0146] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. 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 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0147] 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.

[0148] 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 A step that specifies a function in one or more boxes.

[0149] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0150] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0151] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0152] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity 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, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0153] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0154] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0155] Finally, it should be noted that the method and system for determining a computing power scheduling strategy based on heterogeneous planning disclosed in the embodiments of the present invention are only preferred embodiments of the present invention, and are only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining computing power scheduling strategy based on heterogeneous planning, characterized in that: The method comprises: Get the computing power request set received by multiple network modules in the same computing power network; Determining the request pressure corresponding to each of the network modules according to the request parameters of all the computing power requests of each of the computing power request sets includes: For each computing power request of the computing power request set of each network module, the request parameters of the computing power request are input into a trained load prediction neural network to obtain a request load corresponding to the computing power request; the load prediction neural network is trained by a training data set including a plurality of training computing power request parameters and corresponding load annotations; Calculate the weighted average of the request loads of all the computing power requests in the computing power request set to obtain the request pressure corresponding to the network module; Determining the operation tolerance corresponding to each network module according to the historical operation records and module performance parameters corresponding to each network module, including: For each of the network modules, calculating the average transmission speed of each historical data transmission record in the historical operation record corresponding to the network module, and obtaining the data transmission parameter corresponding to the network module; Calculate the average completion time of all data calculation tasks in the historical data calculation records in the historical operation records corresponding to the network module to obtain the data calculation parameters corresponding to the network module; Calculating a weighted average value of the data transmission parameter and the data calculation parameter to obtain a historical computing capacity parameter of the network module; Inputting the module performance parameters of the network module into a trained computing capacity prediction neural network to obtain the predicted computing capacity parameters corresponding to the network module; the computing capacity prediction neural network is trained by a training data set including a plurality of training module performance parameters and corresponding computing capacity annotations; Calculating a weighted average value of the historical computing capability parameter and the predicted computing capability parameter to obtain a computing tolerance corresponding to the network module; Determine the computing power scheduling strategy corresponding to all the network modules according to the request pressure and the computing tolerance, including: Calculating the ratio of the operation tolerance of each network module to the request pressure to obtain the corresponding load parameters of each network module; The objective function is set to minimize the number of computing power requests allocated to each network module in the allocation scheme; The setting restriction conditions include that the network module with a higher bearing parameter has a higher number of computing power requests allocated, the number of computing power requests allocated to each network module is less than a preset number threshold, and the network module with a bearing parameter lower than a preset parameter threshold does not allocate computing power requests; Based on the dynamic programming algorithm, all computing power requests in the computing power request set are iteratively allocated to each of the network modules according to the objective function and the constraint conditions until the best allocation scheme is obtained to obtain the computing power scheduling strategy.

2. The method for determining computing power scheduling strategy based on heterogeneous planning according to claim 1 is characterized in that: The request parameters include at least one of the requested task data volume, the requested task computing volume, the requested task type, the requested task data type, the requested purpose, the requested task application environment and the requested task programming language.

3. The method for determining computing power scheduling strategy based on heterogeneous planning according to claim 1 is characterized in that: The module performance parameters include hardware device parameters, hardware device models, connection methods between hardware devices, hardware device working coordination methods, and historical hardware device occupancy rates, historical hardware device temperatures, and historical hardware device humidity corresponding to the network module.

4. A computing power scheduling strategy determination system based on heterogeneous planning, characterized in that: The system comprises: An acquisition module is used to obtain a set of computing power requests received by multiple network modules in the same computing power network; The first determination module is used to determine the request pressure corresponding to each of the network modules according to the request parameters of all the computing power requests of each computing power request set, including: For each computing power request of the computing power request set of each network module, the request parameters of the computing power request are input into a trained load prediction neural network to obtain a request load corresponding to the computing power request; the load prediction neural network is trained by a training data set including a plurality of training computing power request parameters and corresponding load annotations; Calculate the weighted average of the request loads of all the computing power requests in the computing power request set to obtain the request pressure corresponding to the network module; The second determination module is used to determine the operation tolerance corresponding to each of the network modules according to the historical operation records and module performance parameters corresponding to each of the network modules, including: For each of the network modules, calculating the average transmission speed of each historical data transmission record in the historical operation record corresponding to the network module, and obtaining the data transmission parameter corresponding to the network module; Calculate the average completion time of all data calculation tasks in the historical data calculation records in the historical operation records corresponding to the network module to obtain the data calculation parameters corresponding to the network module; Calculating a weighted average value of the data transmission parameter and the data calculation parameter to obtain a historical computing capacity parameter of the network module; Inputting the module performance parameters of the network module into a trained computing capacity prediction neural network to obtain the predicted computing capacity parameters corresponding to the network module; the computing capacity prediction neural network is trained by a training data set including a plurality of training module performance parameters and corresponding computing capacity annotations; Calculating a weighted average value of the historical computing capability parameter and the predicted computing capability parameter to obtain a computing tolerance corresponding to the network module; The third determination module is used to determine the computing power scheduling strategies corresponding to all the network modules according to the request pressure and the computing tolerance, including: Calculating the ratio of the operation tolerance of each network module to the request pressure to obtain the corresponding load parameters of each network module; The objective function is set to minimize the number of computing power requests allocated to each network module in the allocation scheme; The setting restriction conditions include that the network module with a higher bearing parameter has a higher number of computing power requests allocated, the number of computing power requests allocated to each network module is less than a preset number threshold, and the network module with a bearing parameter lower than a preset parameter threshold does not allocate computing power requests; Based on the dynamic programming algorithm, all computing power requests in the computing power request set are iteratively allocated to each of the network modules according to the objective function and the constraint conditions until the best allocation scheme is obtained to obtain the computing power scheduling strategy.

5. A computing power scheduling strategy determination system based on heterogeneous planning, characterized in that: The system comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the method for determining a computing power scheduling strategy based on heterogeneous planning as described in any one of claims 1-3.

Images