Training method of scheduling determination model, method and device for determining index data

Abstract

A training method for a scheduling determination model, a method for determining index data, and a device disclosed in an embodiment of the present application, the training method includes obtaining a training data set, the training data set includes multiple pending task sets and a reference index data set corresponding to each pending task set, the reference index data set is a plurality of index data obtained by executing each pending task set based on a reference schedule, the training data set is input into the model to be trained, and a prediction index data set corresponding to each pending task set is obtained, the prediction index data in the prediction index data set corresponds to the reference index data in the reference index data set, and the model to be trained is trained according to the reference index data set and the prediction index data set to obtain a scheduling determination model. The embodiment of the present application selects appropriate scheduling according to the requirements of each pending task in the pending task set, which can reduce congestion and reduce the total execution time of the pending task set.

Description

Technical Field

[0001] The present invention relates to the field of task processing technology, and in particular to a training method for a scheduling determination model, and a method and device for determining indicator data. Background Art

[0002] With the rapid development of automotive electronic technology, the demand for automotive electronics is becoming more diversified and complex. The computational complexity of the on-board microcontroller unit (MCU) is getting higher and higher, and the task processing requirements are increasing. This puts higher requirements on the real-time performance and reliability of the MCU, and also indirectly puts higher requirements on the diversity of the MCU.

[0003] At present, automotive electronics is facing the challenges of energy saving, safety and intelligence in the contemporary globalization. The application fields of MCU are diversified. Behind each challenge is the innovation of semiconductors, and the requirements for MCU in various fields are different. Compared with other fields, the requirements for MCU in automotive control systems are more stringent, including speed and performance requirements, as well as requirements to ensure system reliability and safety. At the same time, in order to cope with complex application scenarios, the requirements for MCU multi-core and multi-tasking are becoming more and more, becoming an important development direction for MCU in the future. Summary of the invention

[0004] The embodiments of the present application provide a method for training a scheduling determination model, a method and device for determining indicator data, which can select appropriate scheduling according to the requirements of each task to be processed in the task set to be processed, reduce congestion, reduce the total execution time of the task set to be processed, improve the processing efficiency of the system, and take into account the real-time performance of the system and the utilization rate of the CPU.

[0005] The present application embodiment provides a method for training a scheduling determination model, including:

[0006] Obtain a training data set; the training data set includes multiple task sets to be processed and a reference indicator data set corresponding to each task set to be processed, and the reference indicator data set is a plurality of indicator data obtained by executing each task set to be processed based on the reference scheduling;

[0007] The training data set is input into the model to be trained to obtain a prediction indicator data set corresponding to each task set to be processed; the prediction indicator data in the prediction indicator data set corresponds one to one with the reference indicator data in the reference indicator data set;

[0008] The model to be trained is trained according to the reference indicator data set and the predicted indicator data set to obtain a scheduling determination model.

[0009] Furthermore, the training data set is input into the model to be trained to obtain the prediction indicator data set corresponding to each task set to be processed, including:

[0010] Determine the priority information and time identification information of each task to be processed in each task set to be processed;

[0011] According to the priority information of each task to be processed and the time identification information of each task to be processed, each task set to be processed is executed based on each candidate schedule in the candidate schedule set to obtain a prediction indicator data set corresponding to each task set to be processed;

[0012] The candidate scheduling in the candidate scheduling set corresponds to the prediction indicator data subset in the prediction indicator data set, and the prediction indicator data in the prediction indicator data subset corresponds one-to-one to the reference indicator data in the reference indicator data set.

[0013] Furthermore, before the model to be trained is trained based on the reference indicator data set and the prediction indicator data set to obtain the scheduling determination model, the following steps are also included:

[0014] According to the priority information of each task to be processed, the weight information of each prediction indicator data in the prediction indicator data set is determined.

[0015] Furthermore, the model to be trained is trained according to the reference indicator data set and the prediction indicator data set to obtain a scheduling determination model, including:

[0016] Determine first evaluation information of the prediction indicator data set and second evaluation information of the reference indicator data set according to weight information of each prediction indicator data in the prediction indicator data set;

[0017] determining error information based on the first evaluation information and the second evaluation information;

[0018] The model to be trained is trained based on the error information to obtain a scheduling determination model.

[0019] Furthermore, the prediction indicator data set includes predicted response time information, predicted execution time information, predicted interrupt frequency information and processor utilization information corresponding to each task to be processed in the task set to be processed.

[0020] The present application embodiment provides a training device for a scheduling determination model, including:

[0021] The first acquisition module is used to acquire a training data set; the training data set includes multiple task sets to be processed and a reference indicator data set corresponding to each task set to be processed, and the reference indicator data set is a plurality of indicator data obtained by executing each task set to be processed based on the reference scheduling;

[0022] The first input-output module is used to input the training data set into the model to be trained to obtain the prediction index data set corresponding to each task set to be processed; the prediction index data in the prediction index data set corresponds to the reference index data in the reference index data set one by one;

[0023] The training module is used to train the model to be trained according to the reference indicator data set and the prediction indicator data set to obtain the scheduling determination model.

[0024] Furthermore, the first input-output module includes:

[0025] A first determination submodule is used to determine the priority information and time identification information of each task to be processed in each task set to be processed;

[0026] The second determination submodule is used to execute each task set to be processed based on each candidate schedule in the candidate schedule set according to the priority information of each task to be processed and the time identification information of each task to be processed, so as to obtain a prediction indicator data set corresponding to each task set to be processed;

[0027] The candidate scheduling in the candidate scheduling set corresponds to the prediction indicator data subset in the prediction indicator data set, and the prediction indicator data in the prediction indicator data subset corresponds one-to-one to the reference indicator data in the reference indicator data set.

[0028] Furthermore, the training device of the scheduling determination model further includes:

[0029] The first determination module is used to determine the weight information of each prediction indicator data in the prediction indicator data set according to the priority information of each task to be processed before training the training model according to the reference indicator data set and the prediction indicator data set to obtain the scheduling determination model.

[0030] Furthermore, the training module includes:

[0031] A first determination submodule, used to determine first evaluation information of the prediction indicator data set and second evaluation information of the reference indicator data set according to weight information of each prediction indicator data in the prediction indicator data set;

[0032] A second determination submodule, used to determine error information according to the first evaluation information and the second evaluation information;

[0033] The training submodule is used to train the model to be trained based on the error information to obtain a scheduling determination model.

[0034] Furthermore, the prediction indicator data set includes predicted response time information, predicted execution time information, predicted interrupt frequency information and processor utilization information corresponding to each task to be processed in the task set to be processed.

[0035] Correspondingly, an embodiment of the present application also provides an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the above-mentioned scheduling determination model training method.

[0036] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the above-mentioned scheduling determination model training method.

[0037] The present application embodiment provides a method for determining indicator data, including:

[0038] Get the set of tasks to be executed;

[0039] The task set to be executed is input into a scheduling determination model, and a target indicator data set corresponding to the task set to be executed is output; the scheduling determination model is a model obtained based on the training method of the above-mentioned scheduling determination model.

[0040] The present application embodiment provides a device for determining indicator data, including:

[0041] The second acquisition module is used to acquire a set of tasks to be executed;

[0042] The second input-output module is used to input the set of tasks to be executed into a scheduling determination model and output a target indicator data set corresponding to the set of tasks to be executed; the scheduling determination model is a model obtained based on the training method of the above-mentioned scheduling determination model.

[0043] Correspondingly, an embodiment of the present application also provides an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the above-mentioned indicator data determination method.

[0044] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the above-mentioned indicator data determination method.

[0045] The embodiments of the present application have the following beneficial effects:

[0046] A method for training a scheduling determination model, a method for determining indicator data, and a device are disclosed in an embodiment of the present application. The training method includes obtaining a training data set, which includes multiple task sets to be processed and a reference indicator data set corresponding to each task set to be processed. The reference indicator data set is a plurality of indicator data obtained by executing each task set to be processed based on a reference schedule, and then the training data set is input into the model to be trained to obtain a prediction indicator data set corresponding to each task set to be processed, wherein the prediction indicator data in the prediction indicator data set corresponds one-to-one to the reference indicator data in the reference indicator data set, and then the model to be trained is trained according to the reference indicator data set and the prediction indicator data set to obtain a scheduling determination model. Based on the embodiment of the present application, a suitable schedule is selected according to the requirements of each task to be processed in the task set to be processed, which can reduce congestion, reduce the total execution time of the task set to be processed, improve the processing efficiency of the system, and take into account the real-time performance of the system and the utilization rate of the CPU. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 It is a schematic diagram of an application scenario of a training method for a scheduling determination model provided in an embodiment of the present application;

[0049] Figure 2 It is a flowchart of a training method for a scheduling determination model provided in an embodiment of the present application;

[0050] Figure 3 It is a structural schematic diagram of a training device for a scheduling determination model provided in an embodiment of the present application;

[0051] Figure 4 This is a schematic diagram of an application scenario of a method for determining indicator data provided in an embodiment of the present application;

[0052] Figure 5 It is a flowchart of a method for determining indicator data provided by an embodiment of the present application;

[0053] Figure 6 It is a structural diagram of an indicator data determination device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical scheme and advantages of the present application clearer, the embodiments of the present application will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiment is only one embodiment of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0055] "Embodiment" referred to herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present application. In the description of the embodiment of the present application, it should be understood that the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the feature defined as "first" and "second" may include one or more of the features explicitly or implicitly. Moreover, the terms "first", "second" and the like are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiment of the present application described here can be implemented in an order other than illustrated or described here. In addition, the terms "including", "having" and "for" and any of their variations are intended to cover non-exclusive inclusions.

[0056] The following describes a specific embodiment of a method for training a scheduling determination model of the present application. Figure 1 is a schematic diagram of an application scenario of a training method for a scheduling determination model provided in an embodiment of the present application, Figure 2 It is a flowchart of a method for training a scheduling determination model provided in an embodiment of the present application. This specification provides method operation steps as shown in the embodiment or flowchart, but may include more or fewer operation steps based on conventional or non-creative labor. The order of steps listed in the embodiment is only one of many execution orders and does not represent the only execution order. In actual execution, the method sequence shown in the embodiment or the accompanying drawings may be executed or executed in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0057] Specific as Figure 2 As shown, the method may include:

[0058] S201: Acquire a training data set; the training data set includes multiple task sets to be processed and a reference indicator data set corresponding to each task set to be processed, and the reference indicator data set is a plurality of indicator data obtained by executing each task set to be processed based on a reference schedule.

[0059] In an embodiment of the present application, the historical operation data of the microcontroller unit (MCU) can be obtained based on the Trace interface of the automotive open system architecture (AUTOSAR) as a training data set. The training data set may include a plurality of task sets to be processed and a reference indicator data set corresponding to each task set to be processed. Optionally, the reference indicator data set may include the response time from ready to execution of the historical task, the net execution time of the historical task, the frequency of interrupt processing of the historical task, and the total execution time of the historical task.

[0060] S203: Input the training data set into the model to be trained to obtain a prediction indicator data set corresponding to each task set to be processed; the prediction indicator data in the prediction indicator data set corresponds one-to-one to the reference indicator data in the reference indicator data set.

[0061] In the embodiment of the present application, after obtaining the training data set, the priority information and time identification information of each task to be processed in each task set to be processed can be determined, and then, according to the priority information and time identification information of each task to be processed in each task set to be processed, each task set to be processed is executed based on each candidate schedule in the candidate schedule set to obtain the prediction index data set corresponding to each task set to be processed. Among them, the candidate schedules in the candidate schedule set correspond to the prediction index data subset in the prediction index data set, and the prediction index data in the prediction index data subset corresponds one-to-one to the reference index data in the reference index data set.

[0062] In an embodiment of the present application, the module to be trained may be a recurrent neural network model.

[0063] In an optional implementation, the priority level of each pending task in the pending task set and the time of arrival at the MCU can be determined, wherein the priority level of each pending task is determined based on the type of each pending task. For example, under normal circumstances, the priority of the pending task corresponding to the on-board safety controller is higher than the priority of the pending task corresponding to the on-board entertainment audio and video controller. The priority of the pending task corresponding to the on-board brake pedal is higher than the priority of the pending task corresponding to the on-board accelerator pedal. Then, according to the priority level of each pending task and the time of arrival at the MCU, each candidate scheduling in the candidate scheduling set can be used to execute the pending task set. Among them, the candidate scheduling set includes but is not limited to first-come-first-served scheduling, short job priority scheduling, time slice round-robin scheduling, priority scheduling, and completely fair scheduling. That is, the pending task set can be executed based on first-come-first-served scheduling, short job priority scheduling, time slice round-robin scheduling, priority scheduling, and completely fair scheduling, and the prediction index data set corresponding to the pending task set under each candidate scheduling is obtained, and the prediction index data set may include but is not limited to the predicted response time information, predicted execution time information, predicted interrupt frequency information, and processor utilization information corresponding to each pending task. Here, although each pending task in the pending task set has a priority level, during the execution of the candidate scheduling, it is not processed according to the priority level of each pending task in the pending task set, but is executed based on the focus of each candidate scheduling. Among them, some schedules focus on real-time performance, but it is difficult to avoid starvation.

[0064] The following example illustrates how to determine the prediction indicator data set corresponding to each task set to be processed.

[0065] Assume that task 1 is a pending task corresponding to the vehicle brake pedal, the priority of task 1 is the highest level L1, and the time when task 1 arrives at MCU is t1; task 2 is a pending task corresponding to the vehicle music player, the priority of task 2 is the lowest level L3, and the time when task 2 arrives at MCU is t2; task 3 is a pending task corresponding to the vehicle positioning device, the priority of task 3 is the medium-high level L2, and the time when task 3 arrives at MCU is t3, where t1 is earlier than t2, and t2 is earlier than t3. The reference scheduling is a first-come-first-served scheduling. Based on the first-come-first-served scheduling, the pending task set including task 1, task 2 and task 3 is executed, and the response time of task 1 is 2s, the execution time is 10s, the interruption is "1 time", and the utilization rate is "70%". The response time of task 2 is 2s, the execution time is 20s, the interruption is "0 times", and the utilization rate is "80%". The response time of task 3 is 4s, the execution time is 30s, the interruption is "0 times", and the utilization rate is "60%". In the model to be trained, the set of pending tasks including Task 1, Task 2 and Task 3 can be executed based on priority scheduling, and the response time of Task 1 is 0s, the execution time is 10s, the interruption is "0 times", and the utilization rate is "60%". The response time of Task 2 is 1s, the execution time is 10s, the interruption is "0 times", and the utilization rate is "70%". The response time of Task 3 is 2s, the execution time is 10s, the interruption is "0 times", and the utilization rate is "80%". In this way, the reference indicator data set can be obtained including {first come first served scheduling│Task 1-response 2s-execution 10s-interruption 1 time-utilization rate 70%, Task 2-response 2s-execution 20s-interruption 0 times-utilization rate 80%, Task 3-response 4s-execution 30s-interruption 0 times-utilization rate 60%}. A subset of prediction index data can also be obtained {priority scheduling | task 1 - response 0s - execution 10s - interruption 0 times - utilization 60%, task 2 - response 1s - execution 10s - interruption 0 times - utilization 70%, task 3 - response 2s - execution 10s - interruption 0 times - utilization 80%}. Similarly, the task set to be processed including task 1, task 2 and task 3 can be executed based on the other candidate scheduling described above to obtain the prediction index data set corresponding to each task set to be processed.

[0066] S205: Train the model to be trained according to the reference indicator data set and the prediction indicator data set to obtain a scheduling determination model.

[0067] In an embodiment of the present application, each task to be processed in the task set to be processed contains different functional modules and has different real-time, priority and other requirements. Therefore, after determining the prediction indicator data set, the weight information of each prediction indicator data in the prediction indicator data set can be determined according to the priority information of each task to be processed, that is, the importance of the prediction indicator data to the trained model to be trained.

[0068] In the embodiment of the present application, since the prediction index data in the prediction index data set corresponds to the reference index data in the reference index data set one by one, the weight information corresponding to each prediction index data in the prediction index data set is the weight information corresponding to the corresponding reference index data. Therefore, the first evaluation information of the prediction index data set and the second evaluation information of the reference index data set can be determined according to the weight information of each prediction index data in the prediction index data set, and the error information can be determined according to the first evaluation information and the second evaluation information, and then the training model is trained based on the error information to obtain the scheduling determination model.

[0069] Based on the examples listed above, let's continue to explain. Assume that the weight corresponding to the response time in the prediction data set corresponding to task 1 is 0.4, the weight corresponding to the execution time is 0.4, the weight corresponding to the interruption is 0.1, and the weight corresponding to the utilization is 0.1. The weight corresponding to the response time in the prediction data set corresponding to task 2 is 0.2, the weight corresponding to the execution time is 0.2, the weight corresponding to the interruption is 0.4, and the weight corresponding to the utilization is 0.2. The weight corresponding to the response time in the prediction data set corresponding to task 3 is 0.3, the weight corresponding to the execution time is 0.3, the weight corresponding to the interruption is 0.2, and the weight corresponding to the utilization is 0.2. At the same time, the preset evaluation rule is that in the one-to-one corresponding reference indicator data and prediction indicator data, the score corresponding to the indicator data with better performance is "1 point", and the score corresponding to the indicator data with poor performance is "0 point". The reference indicator data set {First Come First Served Scheduling│Task 1-Response 2s-Execution 10s-Interruption 1-Utilization 70%, Task 2-Response 2s-Execution 20s-Interruption 0-Utilization 80%, Task 3-Response 4s-Execution 30s-Interruption 0-Utilization 60%} and the prediction indicator data subset {Priority Scheduling│Task 1-Response 0s-Execution 10s-Interruption 0-Utilization 60%, Task 2-Response 1s-Execution 10s-Interruption 0-Utilization 70%, Task 3-Response 2s-Execution 10s-Interruption 0-Utilization 80%} in the above text are used as examples for explanation. The response time of Task 1 in First Come First Served Scheduling is longer than the response time of Task 1 in Priority Scheduling. Therefore, the prediction indicator data "Task 1-Response 0s" gets "1×0.3 points", and the reference indicator data "Task 1-Response 2s" gets "0 points". The execution time of Task 1 in the first-come-first-served scheduling is equal to the response time of Task 1 in the priority scheduling. Therefore, the prediction indicator data "Task 1-Response 10s" gets "1 point", and the reference indicator data "Task 1-Response 10s" gets "1 point". Based on the one-to-one comparison method, the first evaluation information corresponding to the reference indicator data set is "1.3 points", and the second evaluation information corresponding to the prediction indicator data subset is "2.7 points". It can be seen that the second evaluation information corresponding to the prediction indicator data subset is higher than the first evaluation information corresponding to the reference indicator data set, that is, it is more appropriate to execute the pending task set including Task 1, Task 2 and Task 3 based on priority scheduling than based on first-come-first-served scheduling. Based on the same method, the first evaluation information of the prediction indicator data subset corresponding to each candidate scheduling can be determined and iteratively optimized, which will not be repeated here.

[0070] The scheduling determination model obtained by the training method of the scheduling determination model provided in the embodiment of the present application can select appropriate scheduling according to the requirements of each task to be processed in the task set to be processed, which can reduce congestion, reduce the total execution time of the task set to be processed, improve the processing efficiency of the system, and take into account the real-time performance of the system and the utilization rate of the CPU.

[0071] The embodiment of the present application also provides a training device for a scheduling determination model, Figure 3 is a structural diagram of a training device for a scheduling determination model provided in an embodiment of the present application, such as Figure 3 As shown, the device may include:

[0072] The first acquisition module 301 is used to acquire a training data set; the training data set includes a plurality of task sets to be processed and a reference indicator data set corresponding to each task set to be processed, and the reference indicator data set is a plurality of indicator data obtained by executing each task set to be processed based on a reference schedule;

[0073] The first input-output module 303 is used to input the training data set into the model to be trained to obtain a prediction index data set corresponding to each task set to be processed; the prediction index data in the prediction index data set corresponds to the reference index data in the reference index data set one by one;

[0074] The training module 305 is used to train the model to be trained according to the reference indicator data set and the prediction indicator data set to obtain a scheduling determination model.

[0075] Furthermore, the first input-output module 303 may include:

[0076] A first determination submodule is used to determine the priority information and time identification information of each task to be processed in each task set to be processed;

[0077] The second determination submodule is used to execute each task set to be processed based on each candidate schedule in the candidate schedule set according to the priority information of each task to be processed and the time identification information of each task to be processed, so as to obtain a prediction indicator data set corresponding to each task set to be processed;

[0078] The candidate scheduling in the candidate scheduling set corresponds to the prediction indicator data subset in the prediction indicator data set, and the prediction indicator data in the prediction indicator data subset corresponds one-to-one to the reference indicator data in the reference indicator data set.

[0079] Furthermore, the training device of the scheduling determination model further includes:

[0080] The first determination module is used to determine the weight information of each prediction indicator data in the prediction indicator data set according to the priority information of each task to be processed before training the training model according to the reference indicator data set and the prediction indicator data set to obtain the scheduling determination model.

[0081] Furthermore, the training module 305 may include:

[0082] A first determination submodule, used to determine first evaluation information of the prediction indicator data set and second evaluation information of the reference indicator data set according to weight information of each prediction indicator data in the prediction indicator data set;

[0083] A second determination submodule, used to determine error information according to the first evaluation information and the second evaluation information;

[0084] The training submodule is used to train the model to be trained based on the error information to obtain a scheduling determination model.

[0085] Furthermore, the prediction indicator data set includes predicted response time information, predicted execution time information, predicted interrupt frequency information and processor utilization information corresponding to each task to be processed in the task set to be processed.

[0086] The device and method embodiments in the embodiments of the present application are based on the same application concept.

[0087] The training device using the scheduling determination model provided in the embodiment of the present application can select appropriate scheduling according to the requirements of each task to be processed in the task set to be processed, which can reduce congestion, reduce the total execution time of the task set to be processed, improve the processing efficiency of the system, and take into account the real-time performance of the system and the utilization rate of the CPU.

[0088] An embodiment of the present application also provides an electronic device, which can be set in a server to store at least one instruction, at least one program, code set or instruction set related to a training method for a scheduling determination model in an embodiment of the method. The at least one instruction, the at least one program, the code set or instruction set is loaded and executed by the memory to implement the above-mentioned training method for the scheduling determination model.

[0089] An embodiment of the present application also provides a storage medium, which can be set in a server to store at least one instruction, at least one program, code set or instruction set related to a training method for a scheduling determination model in a method embodiment. The at least one instruction, the at least one program, the code set or instruction set is loaded and executed by the processor to implement the above-mentioned scheduling determination model training method.

[0090] Optionally, in this embodiment, the storage medium may be located in at least one of the multiple network servers of the computer network. Optionally, in this embodiment, the storage medium may include, but is not limited to, various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a mobile hard disk, a magnetic disk, or an optical disk.

[0091] The following is a specific embodiment of a method for determining indicator data of the present application. Figure 4 is a schematic diagram of an application scenario of a method for determining indicator data provided in an embodiment of the present application, Figure 5 It is a flowchart of a method for determining indicator data provided by an embodiment of the present application. This specification provides method operation steps as shown in the embodiment or flowchart, but more or fewer operation steps may be included based on conventional or non-creative labor. The order of steps listed in the embodiment is only one of many execution orders and does not represent the only execution order. In actual execution, the method sequence shown in the embodiment or the accompanying drawings may be executed or executed in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0092] Specific as Figure 5 As shown, the method for determining the indicator data may include:

[0093] S501: Obtain a set of tasks to be executed;

[0094] S503: Input the set of tasks to be executed into a scheduling determination model, and output a target indicator data set corresponding to the set of tasks to be executed; the scheduling determination model is a model obtained based on the training method of the above-mentioned scheduling determination model.

[0095] By adopting the method for determining the indicator data provided in the embodiment of the present application, appropriate scheduling can be selected according to the requirements of each task to be processed in the task set to be processed, congestion can be reduced, the total execution time of the task set to be processed can be reduced, the processing efficiency of the system can be improved, and the real-time performance of the system and the utilization rate of the CPU can be taken into account.

[0096] The embodiment of the present application also provides a device for determining indicator data, Figure 6 is a schematic diagram of a structure of an indicator data determination device provided in an embodiment of the present application, such as Figure 6 As shown, the device may include:

[0097] The second acquisition module 601 is used to acquire a set of tasks to be executed;

[0098] The second input-output module 603 is used to input the set of tasks to be executed into a scheduling determination model, and output a target indicator data set corresponding to the set of tasks to be executed; the scheduling determination model is a model obtained based on the training method of the above-mentioned scheduling determination model.

[0099] The device and method embodiments in the embodiments of the present application are based on the same application concept.

[0100] By adopting the method for determining the indicator data provided in the embodiment of the present application, appropriate scheduling can be selected according to the requirements of each task to be processed in the task set to be processed, congestion can be reduced, the total execution time of the task set to be processed can be reduced, the processing efficiency of the system can be improved, and the real-time performance of the system and the utilization rate of the CPU can be taken into account.

[0101] An embodiment of the present application also provides an electronic device, which can be set in a server to store at least one instruction, at least one program, code set or instruction set related to a method for determining indicator data in an implementation method embodiment. The at least one instruction, the at least one program, the code set or instruction set is loaded and executed by the memory to implement the above-mentioned method for determining indicator data.

[0102] An embodiment of the present application also provides a storage medium, which can be set in a server to store at least one instruction, at least one program, code set or instruction set related to a method for determining indicator data in a method embodiment. The at least one instruction, the at least one program, the code set or instruction set is loaded and executed by the processor to implement the above-mentioned method for determining indicator data.

[0103] Optionally, in this embodiment, the storage medium may be located in at least one of the multiple network servers of the computer network. Optionally, in this embodiment, the storage medium may include, but is not limited to, various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a mobile hard disk, a magnetic disk, or an optical disk.

[0104] It can be seen from the embodiments of the training method of the scheduling determination model, the determination method of indicator data, the device, the electronic device or the storage medium provided by the above-mentioned present application that the training method in the present application includes obtaining a training data set, which includes multiple task sets to be processed and a reference indicator data set corresponding to each task set to be processed, and the reference indicator data set is a plurality of indicator data obtained by executing each task set to be processed based on the reference scheduling, and then the training data set is input into the model to be trained to obtain a prediction indicator data set corresponding to each task set to be processed, wherein the prediction indicator data in the prediction indicator data set corresponds to the reference indicator data in the reference indicator data set, and then the model to be trained is trained according to the reference indicator data set and the prediction indicator data set to obtain a scheduling determination model. Based on the embodiments of the present application, appropriate scheduling is selected according to the requirements of each task to be processed in the task set to be processed, which can reduce congestion, reduce the total execution time of the task set to be processed, improve the processing efficiency of the system, and take into account the real-time performance of the system and the utilization rate of the CPU.

[0105] In the present invention, unless otherwise clearly specified and limited, the terms "connected", "connection" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be a connection between two elements or an interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0106] It should be noted that the order of the above embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. The above description describes specific embodiments, and other embodiments are also within the scope of the attached claims. In some cases, the actions or steps recorded in the claims can be performed in the order of different embodiments and can achieve the expected results. In addition, the processes depicted in the drawings do not necessarily require a specific order or connection order to achieve the desired results. In some embodiments, multi-task parallel processing is also possible or may be advantageous.

[0107] 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 device embodiment, since it is based on a method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0108] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for training a scheduling determination model, characterized in that: include: Get the training dataset; The training data set includes a plurality of task sets to be processed and a reference indicator data set corresponding to each task set to be processed, wherein the reference indicator data set is a plurality of indicator data obtained by executing each task set to be processed based on a reference schedule; Determine the priority information and time identification information of each task to be processed in each set of tasks to be processed; According to the priority information of each task to be processed and the time identification information of each task to be processed, each task set to be processed is executed based on each candidate schedule in the candidate schedule set to obtain a prediction indicator data set corresponding to each task set to be processed; wherein the candidate schedules in the candidate schedule set correspond to a prediction indicator data subset in the prediction indicator data set, and the prediction indicator data in the prediction indicator data subset corresponds one-to-one to the reference indicator data in the reference indicator data set; Determining weight information of each prediction indicator data in the prediction indicator data set according to the priority information of each task to be processed; Determining first evaluation information of the prediction indicator dataset and second evaluation information of the reference indicator dataset according to the weight information of each prediction indicator in the prediction indicator dataset; determining error information according to the first evaluation information and the second evaluation information; The model to be trained is trained based on the error information to obtain the scheduling determination model.

2. The method according to claim 1, characterized in that The prediction index data set includes predicted response time information, predicted execution time information, predicted interrupt frequency information and processor utilization information corresponding to each task to be processed in the task set to be processed.

3. A method for determining indicator data, characterized in that: include: Get the set of tasks to be executed; Inputting the set of tasks to be executed into a scheduling determination model, and outputting a target indicator data set corresponding to the set of tasks to be executed; The scheduling determination model is a model obtained based on the training method of the scheduling determination model described in any one of claims 1-2.

4. A training device for a scheduling determination model, characterized in that: include: A first acquisition module is used to acquire a training data set; The training data set includes a plurality of task sets to be processed and a reference indicator data set corresponding to each task set to be processed, wherein the reference indicator data set is a plurality of indicator data obtained by executing each task set to be processed based on a reference schedule; A first input-output module, used to determine the priority information and time identification information of each task to be processed in each set of tasks to be processed; According to the priority information of each task to be processed and the time identification information of each task to be processed, each task set to be processed is executed based on each candidate schedule in the candidate schedule set to obtain a prediction indicator data set corresponding to each task set to be processed; wherein the candidate schedules in the candidate schedule set correspond to a prediction indicator data subset in the prediction indicator data set, and the prediction indicator data in the prediction indicator data subset corresponds one-to-one to the reference indicator data in the reference indicator data set; A training module, used to determine weight information of each prediction indicator data in the prediction indicator data set according to the priority information of each task to be processed; Determining first evaluation information of the prediction indicator dataset and second evaluation information of the reference indicator dataset according to the weight information of each prediction indicator in the prediction indicator dataset; determining error information according to the first evaluation information and the second evaluation information; The model to be trained is trained based on the error information to obtain the scheduling determination model.

5. A device for determining indicator data, characterized in that: include: The second acquisition module is used to acquire a set of tasks to be executed; The second input-output module is used to input the set of tasks to be executed into a scheduling determination model and output a target indicator data set corresponding to the set of tasks to be executed; the scheduling determination model is a model obtained by the training method of the scheduling determination model described in any one of claims 1-2.

6. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the training method of the scheduling determination model described in any one of claims 1-2, or loaded and executed to implement the method for determining the indicator data described in claim 3.

7. A computer-readable storage medium, characterized in that: The storage medium stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set are loaded and executed by the processor to implement the training method of the scheduling determination model as described in any one of claims 1-2, or loaded and executed to implement the method for determining the indicator data as described in claim 3.

Images