Task simulation method and device of virtual simulation platform and computer equipment

Abstract

The invention relates to a task simulation method and device of a virtual simulation platform and computer equipment. The method comprises the following steps: a data synchronization management module performs data synchronization of each heterogeneous virtual simulation sub-platform according to a pre-controlled data synchronization frequency, and further needs to perform online evaluation and offline evaluation so as to optimize an autoregressive moving average model and a reinforcement learning model to perform simulation at the next moment. By adopting the method, the efficiency of data synchronization between heterogeneous simulation platforms can be improved.

Description

Technical Field

[0001] The present application relates to the field of simulation technology, and in particular to a task simulation method, device and computer equipment for a virtual simulation platform. Background Art

[0002] With the development of science and technology, virtual simulation technology is widely used in many fields such as industry, medical treatment, and transportation. Traditional single-architecture simulation platforms are difficult to meet the needs, and heterogeneous virtual simulation platforms have emerged. They are composed of a variety of sub-platforms with different hardware, operating systems, and software. In its operation, data synchronization is a key problem. Due to different tasks of each sub-platform, the system load fluctuates violently, the network delay is unstable due to various factors, the packet loss rate often occurs due to interference, and the clock has deviations.

[0003] Traditional task simulation methods use fixed synchronization frequencies and simple transmission mechanisms, which cannot adapt to real-time changes in the system. Resources are wasted when the load is light, and synchronization is difficult when the load is heavy or the network is poor, resulting in deviations in the results, which cannot meet the high-precision and high-real-time requirements of modern simulation. Therefore, new solutions are urgently needed to solve the data synchronization problem and improve the performance of task simulation. Summary of the invention

[0004] Based on this, it is necessary to provide a task simulation method, device and computer equipment for a virtual simulation platform to address the above technical problems.

[0005] A task simulation method for a virtual simulation platform is applied to the virtual simulation platform, wherein the virtual simulation platform includes a plurality of heterogeneous virtual simulation sub-platforms, a data synchronization management module, and a feedback and supervision module. The method includes: The data synchronization management module performs data synchronization of each heterogeneous virtual simulation sub-platform according to the pre-controlled data synchronization frequency; The supervision module records the real-time simulation status information and the historical data of data synchronization at multiple historical simulation moments in each data synchronization, and constructs a static verification data set based on the historical data; Performing online evaluation and offline evaluation on the data synchronization frequency according to the real-time simulation state information and the static verification data set; The control strategy of the data synchronization frequency is optimized according to the results of the online evaluation and the offline evaluation, so as to perform the task simulation at the next moment according to the optimized control strategy.

[0006] In one of the embodiments, the control strategy of the data synchronization frequency is generated based on an autoregressive moving average model and a deep deterministic policy gradient algorithm; The autoregressive moving average model generates predicted simulation state information at the next moment through the simulation state information at the current moment; the deep deterministic policy gradient algorithm determines the optimal data synchronization frequency through the predicted simulation state information and the simulation state information at the current moment; The simulation status information is: ; is the system load indicator, It's network delay. is the packet loss rate, is the clock bias, is the current synchronization frequency.

[0007] In one embodiment, the method further includes: inputting the simulation state information into a pre-trained autoregressive moving average model, and outputting the predicted simulation state information at the next moment as: ; in, Represents the predicted simulation status information, Indicates the simulation status information at the current moment.

[0008] In one embodiment, the simulation state information and the predicted simulation state information As input; define the action space as increasing synchronization frequency, decreasing synchronization frequency and maintaining synchronization frequency; define the reward function based on the simulation state information; use the deep deterministic policy gradient algorithm to select the best action in each time step to adjust the synchronization frequency.

[0009] In one of the embodiments, it also includes: constructing a shared data stream channel, performing offline evaluation on the autoregressive moving average model and the reinforcement learning model of the deep deterministic policy gradient algorithm according to the static verification data set to obtain offline evaluation information; sending the offline evaluation information to the shared data stream channel; parsing the offline evaluation information in the shared data stream channel, and performing online evaluation on the autoregressive moving average model and the reinforcement learning model according to the offline evaluation information parsing result and the real-time simulation status information.

[0010] In one of the embodiments, it also includes: parsing the offline evaluation information in the shared data stream channel to obtain a load error indicator, a network delay error indicator, a packet loss rate error indicator, and a clock deviation; and performing an online evaluation of the autoregressive moving average model and the reinforcement learning model based on the load error indicator, the network delay error indicator, the packet loss rate error indicator, the clock deviation, and the real-time simulation status information.

[0011] In one of the embodiments, the clock deviation is also monitored in real time, and when the clock deviation is greater than a threshold, the data synchronization frequency is regenerated.

[0012] A task simulation device for a virtual simulation platform is applied to the virtual simulation platform. The virtual simulation platform includes a plurality of heterogeneous virtual simulation sub-platforms, a data synchronization management module, and a feedback and supervision module. The device includes: A synchronization frequency adjustment module, used to synchronize data of various heterogeneous virtual simulation sub-platforms according to a pre-controlled data synchronization frequency; A verification module is used to supervise the module to record the real-time simulation state information and the historical data of data synchronization at multiple historical simulation moments in each data synchronization, and to construct a static verification data set based on the historical data; and to perform online and offline evaluation on the data synchronization frequency based on the real-time simulation state information and the static verification data set; The simulation module is used to optimize the control strategy of the data synchronization frequency according to the results of the online evaluation and the offline evaluation, so as to perform task simulation at the next moment according to the optimized control strategy.

[0013] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented: The data synchronization management module performs data synchronization of each heterogeneous virtual simulation sub-platform according to the pre-controlled data synchronization frequency; The supervision module records the real-time simulation status information and the historical data of data synchronization at multiple historical simulation moments in each data synchronization, and constructs a static verification data set based on the historical data; Performing online evaluation and offline evaluation on the data synchronization frequency according to the real-time simulation state information and the static verification data set; The control strategy of the data synchronization frequency is optimized according to the results of the online evaluation and the offline evaluation, so as to perform the task simulation at the next moment according to the optimized control strategy.

[0014] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps: The data synchronization management module performs data synchronization of each heterogeneous virtual simulation sub-platform according to the pre-controlled data synchronization frequency; The supervision module records the real-time simulation status information and the historical data of data synchronization at multiple historical simulation moments in each data synchronization, and constructs a static verification data set based on the historical data; Performing online evaluation and offline evaluation on the data synchronization frequency according to the real-time simulation state information and the static verification data set; The control strategy of the data synchronization frequency is optimized according to the results of the online evaluation and the offline evaluation, so as to perform the task simulation at the next moment according to the optimized control strategy.

[0015] The task simulation method, device, computer equipment and storage medium of the virtual simulation platform above dynamically adjust the data synchronization frequency through real-time status, thereby realizing adaptive optimization synchronization strategy. However, the synchronization frequency adjustment of multiple heterogeneous virtual simulation sub-platforms is prone to unstable performance after multiple time steps. The performance benchmark of the control strategy of the data synchronization frequency is further verified offline, and then the effectiveness of the control strategy of the data synchronization frequency is determined during online verification, thereby forming a closed loop as a whole. It ensures that efficient and reliable data synchronization services are provided in a complex simulation environment, so as to carry out the tasks of the virtual simulation platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a process flow of a task simulation method of a virtual simulation platform in one embodiment; Figure 2 A structural block diagram of a task simulation device of a virtual simulation platform in an embodiment; Figure 3 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0018] In one embodiment, Figure 1 As shown, a task simulation method of a virtual simulation platform is provided, and the method is applied to the virtual simulation platform, wherein the virtual simulation platform includes a plurality of heterogeneous virtual simulation sub-platforms, a data synchronization management module, and a feedback and supervision module, and includes the following steps: Step 102: The data synchronization management module receives the simulation data of each heterogeneous virtual simulation sub-platform, and collects the simulation status information of each heterogeneous virtual simulation sub-platform in real time through the feedback and supervision module.

[0019] The data synchronization management module includes a central coordination unit, a local control unit, and a data transmission module. The central coordination unit, as a global controller, is responsible for dynamically adjusting the synchronization frequency according to the status of each simulation system (load, delay, clock deviation, etc.). Local control unit: Each simulation system (such as HIL simulation system, VR simulation system, and computer virtual simulation system) has its own local control unit to communicate with the central coordination unit. The local unit collects real-time data such as system load and network status and sends it to the central unit. Data transmission module: Responsible for transmitting data between systems. The transmission rate and frequency are adjusted according to the instructions of the central coordination unit.

[0020] The feedback and supervision modules include: delay monitoring module, system load monitoring module and clock synchronization module. The delay monitoring module monitors the delay time of each data transmission and feeds back to the central coordination unit. The system load monitoring module monitors the CPU, memory, network bandwidth and other load conditions of each simulation system. The clock synchronization module will regularly report the clock deviation between the simulation systems to assist in adjusting the synchronization strategy.

[0021] Step 104: The data synchronization management module performs data synchronization of each heterogeneous virtual simulation sub-platform according to a pre-controlled data synchronization frequency.

[0022] In this step, on the one hand, the system load should be considered. If the synchronization frequency is too high, it may increase the burden on the system when the system load is heavy, causing data processing delays or even system freezes, affecting the overall simulation efficiency. For example, when multiple sub-platforms perform complex calculations at the same time and the amount of data is large, too frequent data synchronization requests may cause an imbalance in the allocation of system resources. On the other hand, network conditions are also a key factor. If the network delay is large or the packet loss rate is high, a too high synchronization frequency may cause data transmission failure or incompleteness, affecting the accuracy and reliability of the data.

[0023] Therefore, the data synchronization management module exchanges data between various heterogeneous sub-platforms in an orderly manner according to the pre-controlled data synchronization frequency. At the specified time interval or under specific trigger conditions (determined by the synchronization frequency), the data transmission process is started to accurately send the new data generated by each sub-platform to other sub-platforms in need, thereby ensuring the coordinated operation of the entire virtual simulation platform, so that each sub-platform can perform subsequent simulation operations and processing based on the latest and consistent data, and finally achieve efficient and accurate virtual simulation tasks.

[0024] Step 106, the supervision module records the real-time simulation state information and the historical data of the data synchronization at multiple historical simulation moments in each data synchronization, and constructs a static verification data set based on the historical data; Step 108: Perform online and offline evaluation on the data synchronization frequency according to the real-time simulation state information and the static verification data set; The performance benchmark of the control strategy of the data synchronization frequency is verified offline, and then the effectiveness of the control strategy of the data synchronization frequency is determined during online verification, thereby forming a closed loop as a whole.

[0025] Step 110: Optimize the control strategy of the data synchronization frequency according to the results of the online evaluation and the offline evaluation, so as to perform task simulation at the next moment according to the optimized control strategy.

[0026] In the task simulation method of the virtual simulation platform, the data synchronization frequency is dynamically adjusted by real-time status to realize the adaptive optimization synchronization strategy. However, the synchronization frequency adjustment of multiple heterogeneous virtual simulation sub-platforms is prone to unstable performance after multiple time steps. The performance benchmark of the control strategy of the data synchronization frequency is further verified offline, and then the effectiveness of the control strategy of the data synchronization frequency is determined during online verification, so as to form a closed loop as a whole. It ensures that efficient and reliable data synchronization services are provided in complex simulation environments, so as to carry out the tasks of the virtual simulation platform.

[0027] In one embodiment, the control strategy of the data synchronization frequency is generated based on an autoregressive moving average model and a deep deterministic policy gradient algorithm; wherein the autoregressive moving average model generates predicted simulation state information at the next moment through simulation state information at the current moment; the deep deterministic policy gradient algorithm determines the optimal data synchronization frequency through the predicted simulation state information and the simulation state information at the current moment; the simulation state information is: ; is the system load indicator, It's network delay. is the packet loss rate, is the clock bias, is the current synchronization frequency.

[0028] Specifically, the autoregressive moving average model undertakes the important task of predicting the simulation state information at the next moment in the entire control strategy. It analyzes and fits the dynamic changes of the system by obtaining the simulation state information at the current moment. Specifically, the autoregressive moving average model comprehensively considers the relationship and change trend of various factors in the system at the current moment, thereby generating a prediction of the simulation state information at the next moment. For example, the autoregressive moving average model will analyze the value, change trend and correlation between the system load index (Lt) at the current moment and other factors (such as network delay Dt, packet loss rate Pt, clock deviation Ct and current synchronization frequency Ft), and predict the possible value and change of the system load index at the next moment based on this. Similarly, other simulation state information such as network delay, packet loss rate, clock deviation, etc. will also be predicted in a similar way, and finally a complete set of simulation state information predicted at the next moment will be obtained.

[0029] The core function of the deep deterministic policy gradient algorithm is to determine the optimal data synchronization frequency. The algorithm uses the predicted simulation state information generated by the autoregressive moving average model and the actual simulation state information at the current moment as input. It extracts and learns the above information through a deep neural network to build a policy model that can reflect the relationship between the system state and the optimal synchronization frequency. In this model, the algorithm will continuously explore and evaluate the performance of the system under different synchronization frequencies. According to a certain reward mechanism, for example, when the system load is low, the network delay is small, the packet loss rate is low, and the clock deviation is within an acceptable range, a higher reward is given, otherwise a lower reward is given, and the strategy is gradually optimized to finally determine the optimal data synchronization frequency that can achieve the best system performance.

[0030] In another embodiment, the simulation state information is input into a pre-trained autoregressive moving average model, and the predicted simulation state information at the next moment is output as: ; in, Represents the predicted simulation status information, Indicates the simulation status information at the current moment.

[0031] Specifically, the autoregressive moving average model predicts the future system status based on historical data collected in real time: 1. Predict network delay: Predict possible network delay peaks in the future so as to adjust the synchronization frequency in advance 2. Predict system load: Predict the changing trend of system load to ensure that the synchronization frequency is reduced when the load increases to avoid system overload.

[0032] 3. Predict clock deviation: Detect changes in clock deviation and adjust the clock synchronization strategy in advance.

[0033] Based on the results of time series forecasting, the system can adjust the synchronization frequency in advance. For example: 1. If it is predicted that network latency will increase in the future, the synchronization frequency can be reduced in advance to reduce the pressure of data transmission.

[0034] 2. If it is predicted that the system load will decrease in the future, the synchronization frequency can be increased to improve the accuracy and real-time performance of synchronization.

[0035] In one embodiment, the simulation state information and the predicted simulation state information As input; define the action space as increasing synchronization frequency, decreasing synchronization frequency and maintaining synchronization frequency; define the reward function based on the simulation state information; use the deep deterministic policy gradient algorithm to select the best action in each time step to adjust the synchronization frequency.

[0036] Each action corresponds to a specific synchronization frequency adjustment strategy. For example: 1. When increasing the synchronization frequency, data synchronization is adjusted from 10 times per second to 20 times per second.

[0037] 2. When reducing the synchronization frequency, the synchronization frequency is adjusted from 20 times per second to 10 times per second.

[0038] In one of the embodiments, a shared data stream channel is constructed, and an autoregressive moving average model and a reinforcement learning model are evaluated offline based on a static verification data set to obtain offline evaluation information; the offline evaluation information is sent to the shared data stream channel; the offline evaluation information in the shared data stream channel is parsed, and the autoregressive moving average model and the reinforcement learning model are evaluated online based on the offline evaluation information parsing results and real-time simulation status information.

[0039] In an embodiment, a shared data stream channel is constructed to ensure real-time sharing and updating of data during offline and online verification. After obtaining offline evaluation information, a message queue (such as Kafka, RabbitMQ) or a stream processing framework (such as Apache Flink) is used to achieve real-time data transmission to ensure low latency and high throughput. It is worth noting that since data exchange is relatively frequent during task simulation, when setting historical data, the historical data is not updated in real time, but updated at preset intervals, such as setting the update time to 1 hour, etc. At this time, the network environment of multiple heterogeneous virtual simulation sub-platforms will not change much, so the performance standards obtained by offline training can meet the heterogeneous simulation of multiple heterogeneous virtual simulation sub-platforms.

[0040] In one of the embodiments, the offline evaluation information includes: a load error indicator, a network delay error indicator, a packet loss rate error indicator, and a clock deviation; the above-mentioned error indicators can use RMSE and MAE indicators to parse the offline evaluation information in the shared data stream channel to obtain the load error indicator, the network delay error indicator, the packet loss rate error indicator, and the clock deviation error indicator; according to the load error indicator, the network delay error indicator, the packet loss rate error indicator, the clock deviation error indicator, and the real-time simulation status information, the autoregressive moving average model and the reinforcement learning model are evaluated online.

[0041] In this embodiment, the load error index, network delay error index, packet loss rate error index and clock deviation error index are the performance standards of the autoregressive moving average model and the reinforcement learning model realized in this period of history. In the online evaluation stage, the performance standards can be used for online evaluation. When the online evaluation results reach the performance standards, it means that the performance of the autoregressive moving average model and the reinforcement learning model can meet the task simulation, or by setting weights or other methods, the performance standards of the online evaluation are higher than the performance standards of the offline evaluation, so that the autoregressive moving average model and the reinforcement learning model are further optimized in the online evaluation process, or the online evaluation shows that the performance of the autoregressive moving average model and the reinforcement learning model is lower than the performance standards, and the autoregressive moving average model and the reinforcement learning model also need to be further optimized. It is worth noting that through offline and online evaluation, the entire simulation process forms a closed loop, which greatly improves the task simulation reliability of the virtual simulation platform.

[0042] In one of the embodiments, the clock deviation is monitored in real time, and when the clock deviation is greater than a threshold, the clock synchronization is performed by reducing the synchronization frequency until the clock deviation is less than the threshold.

[0043] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0044] In one embodiment, Figure 2As shown, a task simulation device of a virtual simulation platform is provided, which is applied to the virtual simulation platform. The virtual simulation platform includes multiple heterogeneous virtual simulation sub-platforms, a data synchronization management module and a feedback and supervision module, including: a synchronization frequency adjustment module 202, a verification module 204 and a simulation module 206, wherein: A synchronization frequency adjustment module 202, used to synchronize data of various heterogeneous virtual simulation sub-platforms according to a pre-controlled data synchronization frequency; Verification module 204, used for monitoring the module to record the real-time simulation state information and the historical data of data synchronization at multiple historical simulation moments in each data synchronization, and to construct a static verification data set according to the historical data; and to perform online and offline evaluation on the data synchronization frequency according to the real-time simulation state information and the static verification data set; The simulation module 206 is used to optimize the control strategy of the data synchronization frequency according to the results of the online evaluation and the offline evaluation, so as to perform task simulation at the next moment according to the optimized control strategy.

[0045] In one embodiment, the control strategy of the data synchronization frequency is generated based on an autoregressive moving average model and a deep deterministic policy gradient algorithm; wherein the autoregressive moving average model generates predicted simulation state information at the next moment through simulation state information at the current moment; the deep deterministic policy gradient algorithm determines the optimal data synchronization frequency through the predicted simulation state information and the simulation state information at the current moment; the simulation state information is: ; is the system load indicator, It's network delay. is the packet loss rate, is the clock bias, is the current synchronization frequency.

[0046] In one embodiment, the synchronization frequency adjustment module 202 is further used to input the simulation state information into a pre-trained autoregressive moving average model, and output the predicted simulation state information at the next moment as: ; in, Represents the predicted simulation status information, Indicates the simulation status information at the current moment.

[0047] In one embodiment, the synchronization frequency adjustment module 202 is further configured to convert the simulation state information and the predicted simulation state information into As the input of the reinforcement learning model; define the action space as increasing the synchronization frequency, decreasing the synchronization frequency, and maintaining the synchronization frequency; define the reward function based on the simulation state information; use the deep deterministic policy gradient algorithm to select the best action in each time step to adjust the synchronization frequency.

[0048] In one embodiment, the verification module 204 is also used to construct a shared data flow channel, perform offline evaluation on the autoregressive moving average model and the reinforcement learning model of the deep deterministic policy gradient algorithm according to the static verification data set to obtain offline evaluation information; send the offline evaluation information to the shared data flow channel; parse the offline evaluation information in the shared data flow channel, and perform online evaluation on the autoregressive moving average model and the reinforcement learning model according to the offline evaluation information parsing result and the real-time simulation status information.

[0049] In one of the embodiments, the offline evaluation information includes: a load error indicator, a network delay error indicator, a packet loss rate error indicator, and a clock deviation error indicator; the verification module 204 is also used to parse the offline evaluation information in the shared data stream channel to obtain the load error indicator, the network delay error indicator, the packet loss rate error indicator, and the clock deviation; based on the load error indicator, the network delay error indicator, the packet loss rate error indicator, the clock deviation, and the real-time simulation status information, the autoregressive moving average model and the reinforcement learning model are evaluated online.

[0050] In one embodiment, the synchronization frequency adjustment module 202 is further configured to monitor the clock deviation in real time, and regenerate the data synchronization frequency when the clock deviation is greater than a threshold.

[0051] For the specific definition of the task simulation device of the virtual simulation platform, please refer to the definition of the task simulation method of the virtual simulation platform in the above text, which will not be repeated here. The various modules in the task simulation device of the above virtual simulation platform can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0052] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 3As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a task simulation method of a virtual simulation platform is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0053] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0054] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method in the above embodiment when executing the computer program.

[0055] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method in the above embodiment are implemented.

[0056] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0057] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0058] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A task simulation method for a virtual simulation platform, characterized in that: Applied in a virtual simulation platform, the virtual simulation platform includes a plurality of heterogeneous virtual simulation sub-platforms, a data synchronization management module and a feedback and supervision module, the method includes: The data synchronization management module performs data synchronization of each heterogeneous virtual simulation sub-platform according to the pre-controlled data synchronization frequency; The feedback and supervision module records the real-time simulation state information and the historical data of data synchronization at multiple historical simulation moments in each data synchronization, and constructs a static verification data set based on the historical data; Performing online evaluation and offline evaluation on the data synchronization frequency according to the real-time simulation state information and the static verification data set; The control strategy of the data synchronization frequency is optimized according to the results of the online evaluation and the offline evaluation, so as to perform the task simulation at the next moment according to the optimized control strategy.

2. The method according to claim 1, characterized in that The control strategy of the data synchronization frequency is generated based on an autoregressive moving average model and a deep deterministic policy gradient algorithm; The autoregressive moving average model generates predicted simulation state information at the next moment through the simulation state information at the current moment; the deep deterministic policy gradient algorithm determines the optimal data synchronization frequency through the predicted simulation state information and the simulation state information at the current moment; The simulation status information is: ； is the system load indicator, It's network latency. is the packet loss rate, is the clock bias, is the current synchronization frequency.

3. The method according to claim 2, characterized in that The autoregressive moving average model generates predicted simulation state information at the next moment through the simulation state information at the current moment, including: The simulation state information is input into the pre-trained autoregressive moving average model, and the predicted simulation state information at the next moment is output as: ； in, Represents the prediction simulation status information, Indicates the simulation status information at the current moment.

4. The method according to claim 3, characterized in that The deep deterministic policy gradient algorithm determines the optimal data synchronization frequency through the predicted simulation state information and the simulation state information at the current moment, including: The simulation state information and the predicted simulation state information As input; The action space is defined as increasing synchronization frequency, decreasing synchronization frequency, and maintaining synchronization frequency; Define a reward function based on simulation state information; A deep deterministic policy gradient algorithm is used to select the best action at each time step to adjust the synchronization frequency.

5. The method according to claim 4, characterized in that Performing online evaluation and offline evaluation on the data synchronization frequency according to the real-time simulation state information and the static verification data set includes: Constructing a shared data stream channel, and performing offline evaluation on the autoregressive moving average model and the reinforcement learning model corresponding to the deep deterministic policy gradient algorithm according to the static verification data set to obtain offline evaluation information; Sending the offline evaluation information to the shared data stream channel; The offline evaluation information in the shared data stream channel is parsed, and the autoregressive moving average model and the reinforcement learning model are evaluated online according to the offline evaluation information parsing result and the real-time simulation state information.

6. The method according to claim 5, characterized in that The offline evaluation information includes: load error index, network delay error index, packet loss rate error index and clock deviation error index; Parsing the offline evaluation information in the shared data stream channel, and performing online evaluation on the autoregressive moving average model and the reinforcement learning model according to the offline evaluation information parsing result and the real-time simulation state information, including: Parsing the offline evaluation information in the shared data stream channel to obtain a load error indicator, a network delay error indicator, a packet loss rate error indicator, and a clock deviation; The autoregressive moving average model and the reinforcement learning model are evaluated online according to the load error index, the network delay error index, the packet loss rate error index, the clock deviation and the real-time simulation state information.

7. The method according to any one of claims 2 to 6, characterized in that: The method further comprises: The clock deviation is monitored in real time, and when the clock deviation is greater than a threshold, the data synchronization frequency is regenerated.

8. A task simulation device for a virtual simulation platform, characterized in that: Applied in a virtual simulation platform, the virtual simulation platform includes a plurality of heterogeneous virtual simulation sub-platforms, a data synchronization management module and a feedback and supervision module, the device includes: A synchronization frequency adjustment module, used to synchronize data of various heterogeneous virtual simulation sub-platforms according to a pre-controlled data synchronization frequency; The verification module is used for the feedback and supervision module to record the real-time simulation state information and the historical data of data synchronization at multiple historical simulation moments in each data synchronization, and to construct a static verification data set based on the historical data; and to perform online and offline evaluation on the data synchronization frequency based on the real-time simulation state information and the static verification data set; The simulation module is used to optimize the control strategy of the data synchronization frequency according to the results of the online evaluation and the offline evaluation, so as to perform task simulation at the next moment according to the optimized control strategy.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images