CN120104288B

CN120104288B - Scheduling method, equipment and storage medium of semi-physical simulation resources

Info

Publication number: CN120104288B
Application number: CN202510592709.3A
Authority: CN
Inventors: 王宇; 包富瑜; 陈功; 覃金贵; 邬晓毅; 苏金波; 张俊傲; 李航; 钟宇; 叶志强; 王勇
Original assignee: Chengdu Fluid Power Innovation Center
Current assignee: Chengdu Fluid Power Innovation Center
Priority date: 2025-05-09
Filing date: 2025-05-09
Publication date: 2025-07-11
Anticipated expiration: 2045-05-09
Also published as: CN120104288A

Abstract

The present application relates to the field of semi-physical simulation, and in particular to a scheduling method, device and storage medium for semi-physical simulation resources. The method includes: obtaining a number of simulation models, a number of simulation devices and a number of test tasks in a semi-physical simulation resource library; combining the simulation model with different simulation devices to obtain a semi-physical simulation node, and generating a variety of collaborative combinations to collaboratively execute multiple test tasks to obtain performance test data for each simulation model; analyzing the performance test data of each simulation model, generating a performance parameter set for each simulation model, and the performance parameter set includes a number of scene-sensitive indicators and a number of static basic indicators; based on the simulation model requirements and the performance parameter set of each simulation model, selecting a target simulation model and a target simulation device from the semi-physical simulation resource library to execute the target simulation task. A high degree of adaptation of the scheduling strategy to the actual operating environment is achieved, which improves resource utilization and scheduling accuracy.

Description

Scheduling method, equipment and storage medium of semi-physical simulation resources

Technical Field

The present application relates to the field of semi-physical simulation, and in particular, to a method, an apparatus, and a storage medium for scheduling semi-physical simulation resources.

Background

The semi-physical simulation (HIL) technology provides an efficient and reliable means for developing, testing and verifying complex systems by organically combining physical equipment with a digital simulation model, and the application scene of the technology covers the fields of aerospace, automotive electronics, power systems, industrial automation and the like, so that the development cost can be remarkably reduced, and the product marketing period can be shortened. If advanced resource scheduling technology can be combined in the HIL framework, the simulation efficiency and efficiency can be further improved.

For example, the patent application publication number CN116774978a discloses a reusable framework of a service-oriented simulation model and an integration method, the reusable framework of the service-oriented simulation model comprises a resource layer, a service component layer, a service layer and an application layer, the resource layer is used for solving functions of description, organization and management of resources, the service component layer is used for packaging models with different layers, different fields, different categories and different granularities in the form of components, the service component layer mainly comprises component description, component definition, component construction tools, component verification, component configuration, component devices, component reconstruction, component packaging and the like, the main tasks are to complete description, reconstruction and assembly of the models, and the service layer is used for packaging the models in a cloud service mode, so that the isomerism among the simulation service models with different layers, different granularities and different fields can be shielded. For another example, the patent application publication number CN119088562A discloses a simulation resource pool management scheduling method and system, a simulation method and system, wherein the simulation resource pool management scheduling method comprises the steps of constructing a device resource pool and a model resource pool, wherein related information of a plurality of simulation devices is configured in the device resource pool, a plurality of model runtime state information is configured in the model resource pool, a simulation task is established, runtime state information of a model corresponding to the simulation task is acquired from the model resource pool, a priority value matched with the model by the simulation devices in the device resource pool is calculated, the simulation devices are regulated based on the optimal priority value, and the regulated simulation devices are distributed to the corresponding simulation tasks.

However, in order to realize standardized management, the conventional resource scheduling technology often adopts unified data encapsulation, so that the joint scheduling requirement of a simulation model and physical equipment in semi-physical simulation is difficult to meet, the dynamic allocation and elastic scheduling capability of resources are insufficient, and the efficiency and expansibility of a simulation system are limited.

Disclosure of Invention

The application mainly aims to provide a scheduling method, equipment and storage medium of semi-physical simulation resources. In order to solve the technical problems, the application adopts the following technical scheme:

The first aspect of the present application provides a scheduling method of semi-physical simulation resources, the method comprising:

S101, acquiring a semi-physical simulation resource library and test tasks under a plurality of different task scenes, wherein the semi-physical simulation resource library comprises a plurality of simulation models and a plurality of simulation devices;

S102, combining each simulation model with different simulation devices to obtain a plurality of semi-physical simulation nodes, wherein the simulation models or the simulation devices among the plurality of semi-physical simulation nodes are different;

s103, generating a plurality of cooperation combinations based on the plurality of semi-physical simulation nodes to cooperatively execute a plurality of test tasks to obtain performance test data of each simulation model under different simulation equipment, different cooperation combinations and different task scenes;

S104, analyzing performance test data of each simulation model to generate a performance parameter set of each simulation model, wherein the performance parameter set comprises a plurality of scene sensitivity indexes and a plurality of static basic indexes;

S105, selecting a target simulation model and target simulation equipment from the semi-physical simulation resource library according to the simulation model requirements and the performance parameter set of each simulation model based on the task scene and the simulation model requirements of the target simulation task so as to execute the target simulation task.

In some embodiments, the performance test data comprises a plurality of evaluation indexes, S104 comprises the steps of comparing a plurality of values of the same evaluation index in the performance test data of each simulation model, determining value fluctuation, determining corresponding evaluation indexes as scene sensitive indexes when the value fluctuation is larger than a preset fluctuation value, storing different values of the scene sensitive indexes in association with task scenes and/or simulation equipment and/or cooperation combinations, and determining corresponding evaluation indexes as the static basic indexes when the value fluctuation is smaller than the preset fluctuation value.

In some embodiments, the method further comprises the steps of analyzing the change trend of a plurality of values of the scene sensitive index, respectively determining the influence coefficient of the task scene, the simulation equipment and the cooperation combination on the scene sensitive index, determining the influence factor of the scene sensitive index from the task scene, the simulation equipment and the cooperation combination according to the influence coefficient, and storing the influence factor in association with the scene sensitive index.

In some embodiments, the method further comprises storing the static base indicator as a fixed value or a range interval, wherein the range interval is determined based on a plurality of values of the static base indicator.

In some embodiments, the simulation model requirements include performance requirements, the step 105 further includes screening a plurality of unused simulation models based on the performance requirements and the static basic indexes of the simulation models to obtain a plurality of alternative simulation models, generating a plurality of alternative simulation nodes based on the plurality of alternative simulation models and a plurality of unused alternative simulation devices in the semi-physical simulation resource library, obtaining a plurality of alternative collaboration combinations based on the plurality of alternative simulation node combinations, evaluating predicted performance and resource occupancy rate of each alternative collaboration combination in a task scene of the target simulation task based on scene sensitivity indexes of the alternative simulation models, selecting the alternative collaboration combination with the predicted performance meeting the performance requirements and the lowest resource occupancy rate as a target collaboration combination, and taking the alternative simulation models and the alternative simulation devices in the target collaboration combination as the target simulation model and the target simulation device.

In some embodiments, the step S105 further includes obtaining the priority of the simulation task to be executed, and taking at least one simulation task with the highest priority as the target simulation task.

In some embodiments, the method further comprises loading the target simulation model into corresponding target simulation equipment to obtain a plurality of target semi-physical simulation nodes, continuously monitoring task execution quality of the plurality of target semi-physical simulation nodes in the process of executing the target simulation task, acquiring actual measurement performance data of the target semi-physical simulation nodes when the task execution quality is lower than preset quality, and isolating simulation data generated by the target semi-physical simulation nodes if the difference between the actual measurement performance data of the target semi-physical simulation nodes and the performance test data is larger than preset difference.

In some embodiments, the method further comprises selecting a preferred simulation device from a plurality of unused simulation devices in the semi-physical simulation resource library based on a scene sensitivity index of the target simulation model if the difference between the measured performance data and the performance test data of the target semi-physical simulation node is less than the preset difference, loading the target simulation model into the preferred simulation device to replace the target simulation device, and updating the target semi-physical simulation node.

A second aspect of the present application provides a computer apparatus, the apparatus comprising:

a memory for storing a computer program;

And the processor is used for executing the computer program and realizing the steps of the scheduling method of the semi-physical simulation resource provided by any embodiment of the application when the computer program is executed.

The third aspect of the present application also provides a computer readable storage medium, which stores a computer program, the computer program when executed by a processor causes the processor to perform the steps of the scheduling method of semi-physical simulation resources provided by any embodiment of the present application.

The beneficial effects are that:

The embodiment of the application provides a scheduling method, equipment and storage medium of semi-physical simulation resources, wherein a dynamic and static separation index system is established in a performance parameter set of each simulation model through diversified heterogeneous resource test combinations, and model performance is divided into a scene sensitive index and a static basic index and customized storage so as to display the performance boundary and dynamic influence relation of the simulation model in different environments. Therefore, the strong coupling between the semi-physical simulation resources is converted into the advantage of improving the cooperative efficiency, so that the semi-physical simulation resources can be based on multidimensional analysis of the performance parameter set, the high adaptation of the scheduling strategy and the actual running environment is realized, and the resource utilization rate and the scheduling accuracy are improved.

Specifically, the simulation model and different simulation devices are combined in a crossing mode to generate diversified simulation nodes, test tasks are executed under a multi-task scene, complex scenes under different test conditions (such as task scene types, simulation device models and collaborative combination configuration) are covered, and then collaborative efficiency and implicit dependency relations of the digital model, physical hardware and different model combinations are captured. Furthermore, the index which is obviously affected is identified by comparing the performance parameter sets, the dynamic index and the static index are separated, the dynamic index and the key impact factors thereof are associated and stored, the data packaging mode of 'one cut' is broken, and a basis is provided for dynamic scheduling.

Furthermore, in the initial stage of task matching, models which do not meet the basic capability requirement are rapidly filtered through static indexes, so that the computational complexity of subsequent dynamic evaluation is reduced. And quantifying real-time influence of task scenes, equipment characteristics and a cooperation mode on model performance through scene sensitivity indexes, realizing resource scheduling by adopting a layer-by-layer screening and overall screening mechanism based on multidimensional (simulation performance and resource occupation), aiming at performance degradation or complementary effect under a strong coupling environment, pre-judging and avoiding performance traps, breaking through the limitation of single resource performance evaluation, and optimizing the dynamic response capability of resource scheduling on a complex environment.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. Like elements or portions are generally identified by like reference numerals throughout the several figures. In the drawings, elements or portions thereof are not necessarily drawn to scale. It will be apparent to those of ordinary skill in the art that the drawings in the following description are of some embodiments of the application and that other drawings may be derived from these drawings without inventive faculty.

FIG. 1 is a schematic flow chart of a scheduling method of semi-physical simulation resources provided by an embodiment of the application;

FIG. 2 is a schematic diagram of a semi-physical simulation resource library package provided by an embodiment of the present application;

fig. 3 is a schematic block diagram of a computer device according to an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. It will be apparent that the described embodiments are some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

In this document, suffixes such as "module", "component", or "unit" used to represent elements are used only for facilitating the description of the present application, and have no particular meaning in themselves. Thus, "module," "component," or "unit" may be used in combination.

The terms "upper," "lower," "inner," "outer," "front," "rear," "one end," "the other end," and the like herein refer to an orientation or positional relationship based on that shown in the drawings, merely for convenience of description and to simplify the description, and do not denote or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the application. Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.

Herein, "and/or" includes any and all combinations of one or more of the associated listed items.

Herein, "plurality" means two or more, i.e., it includes two, three, four, five, etc.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

Semi-physical simulation is a technique that combines actual physical hardware with a computer-generated digital model for system testing. It verifies the performance and reliability of the system by running real hardware and virtual models in a closed loop environment.

Herein, a simulation model refers to a mathematical or logical model for simulating system behavior, i.e., a virtual model in semi-physical simulation. These models may be kinetic models constructed based on laws of physics, control system algorithms, or signal processing flows, etc. They are typically implemented in simulation software and are capable of generating output results in response to input data.

As used herein, simulation devices refer to those actual physical components, such as sensors, actuators, controllers, etc., that directly participate in the experimental process, i.e., the actual hardware devices in semi-physical simulation. The hardware devices receive data from the simulation model as input and convert the data into actual physical actions or state changes, and meanwhile, the hardware devices can feed back the actual operation results to the simulation model.

In semi-physical simulation systems, there is a close relationship between the simulation model and the simulation equipment. On the one hand, the simulation model provides control instructions or environment parameters for the physical equipment, and on the other hand, the physical equipment executes corresponding operations according to the received information and feeds the results back to the simulation model to update the state of the simulation model. This interaction requires a high degree of consistency and synchronization between the two, and the strong coupling between the simulation model and the physical device also results in interaction between the two. In addition, along with the expansion of the system scale and the diversification of application scenes, simulation models in different fields often need to be combined to work together to form heterogeneous model combined simulation, so that the models can also influence each other. For example, the unmanned plane model completes the observation task, the radar model completes the detection task, the two simulation models cooperate with each other to complete the task, and the performance of the two simulation models has a coupling relationship.

The existing resource scheduling technology tends to adopt a unified data packaging mode, normalizes a data format, is convenient to store and retrieve, but exposes serious limitations when facing cross-equipment and cross-field scenes. Because of the lack of flexible adaptation mechanism, different types of data are difficult to be in seamless connection, information islands are easy to form, even if the information islands have rich resource libraries, dynamic allocation and flexible scheduling of resources cannot be accurately realized, and efficiency and expansibility of a simulation system are limited.

Based on the above, the embodiment of the application provides a scheduling method, equipment and storage medium of semi-physical simulation resources, through diversified heterogeneous resource test combinations, a dynamic and static separation index system is established in a performance parameter set of each simulation model, and model performance is divided into a scene sensitive index and a static basic index and customized storage so as to display the relationship between performance boundaries and dynamic influences of the simulation models in different environments. The static basic index quantifies the inherent performance of the model, is used for screening the simulation model with the basic capability reaching the standard, and the scene sensitivity index dynamically reflects the performance fluctuation and suitability of the model in the real deployment environment and is used for performing the performance evaluation of the joint simulation. Therefore, the strong coupling between semi-physical simulation resources is converted into the advantage of improving the cooperative efficiency, so that the semi-physical simulation resources can be flexibly matched with task requirements and resource characteristics based on multidimensional analysis of performance parameter sets, and performance traps are prejudged and avoided.

Some embodiments of the present application are described in detail below with reference to the accompanying drawings. The following embodiments and features of the embodiments may be combined with each other without conflict. Referring to fig. 1, fig. 1 is a schematic flowchart of a method for scheduling semi-physical simulation resources according to an embodiment of the present application, and as shown in fig. 1, the embodiment of the present application provides a method for scheduling semi-physical simulation resources, which includes S101 to S105.

S101, acquiring a semi-physical simulation resource library and test tasks under a plurality of different task scenes, wherein the semi-physical simulation resource library comprises a plurality of simulation models and a plurality of simulation devices.

The semi-physical simulation resource library is a comprehensive resource pool which stores various simulation models (such as an aircraft model, a radar model, an unmanned aerial vehicle model, a satellite model and the like) and various simulation devices (such as a real-time simulator, a five-axis flight turntable, a moment load simulator, a flight control computer, an external hanging task computer, a GPS satellite navigation generator and the like).

Specifically, multi-dimensional task scenarios are defined to cover combinations of different task types (e.g., probing tasks, collaborative tasks) and different task contexts (e.g., mountain terrain, bad weather conditions, city dense building groups, electromagnetic interference), each scenario corresponding to a particular simulation condition and constraint. Correspondingly, the test task is designed for specific verification requirements of different task scenes, such as testing communication stability of unmanned aerial vehicle formation in rainy environment or evaluating radar detection accuracy in urban canyon scene. By way of example, a limited number of representative task scenarios may be preset according to actual requirements, on one hand, the cooperative suitability of the full-coverage model and the device under various conditions, and on the other hand, the redundant or invalid test tasks are avoided, and the early test burden is increased, so that accurate and reliable basic data is provided for subsequent performance analysis, and specific task scenarios and test tasks are not limited herein.

It should be appreciated that by simulating the complexity of a real scene, the adaptability and collaborative performance of the simulation resources in heterogeneous environments are systematically verified. For example, the detection task may test the perceived accuracy decay of the model in extreme weather, and the cooperative interaction task may test the multi-device communication delay fluctuations in electromagnetic interference. The multi-dimensional test can acquire fine-granularity dynamic performance data, provide differentiated scene adaptation basis for subsequent resource scheduling, and avoid performance evaluation deviation caused by single scene test.

S102, combining each simulation model with different simulation devices to obtain a plurality of semi-physical simulation nodes, wherein the simulation models or the simulation devices among the plurality of semi-physical simulation nodes are different.

Specifically, the semi-physical simulation node refers to a minimum execution unit formed by a simulation model and at least one simulation device, wherein the simulation model is responsible for simulating digital logic of a tested system, the simulation device provides a physical interface or a hardware environment, the simulation model and the simulation device cooperatively complete a specific simulation task, and the simulation model and the simulation device are combined in a full-crossing way to generate the diversified semi-physical simulation node. For example, the aircraft model A can be respectively combined with the flight control equipment and the five-axis flight turntable to obtain two semi-physical simulation nodes.

It should be understood that various semi-physical simulation nodes cover the possibility of diversification of models and devices in hardware-software cooperation, a complete set of 'model-device' pairing is formed, suitability of the models on different devices is tested through the difference between the nodes, and a foundation is provided for subsequent analysis of software-hardware cooperation efficiency. For example, the simulation model can be loaded to the simulation equipment according to the model-equipment pairing in the semi-physical simulation node, and then the simulation model can be used for executing the test task or the simulation task.

In some embodiments, the emulation device may be a single hardware device or may be a hardware system integrated by multiple hardware devices. That is, a plurality of hardware devices may be integrated in each simulation device, and the simulation devices are divided into lightweight nodes and full-scale nodes according to the depth and breadth that the simulation devices can simulate. The lightweight node can be used for simulating or replacing functions of some miniature controllers or sensors in the system, and the full-quantization node is closer to a complete hardware system in practical application, and provides more comprehensive and accurate functional simulation or directly operates as a part of a more reliable system compared with the lightweight node.

In unmanned aerial vehicle simulation, the lightweight node can comprise a real-time simulator, a flight control computer and an externally hung task computer, is used for quick verification of core functions and meets the simulation requirements of lightweight and low cost, the full-quantized node can comprise a real-time simulator, a five-axis flight turntable, a moment load simulator, a flight control computer, an externally hung task computer, a GPS satellite navigation generator and the like, the configuration of the full-quantized node is more comprehensive, and the execution of tasks such as flight, detection and the like in a real environment can be simulated more accurately. The real-time simulator is a core computing unit and is used for running a corresponding simulation model, an environment model and a control algorithm, the flight control computer is used for executing flight control logic (such as attitude control and navigation algorithm), the plug-in mission computer is used for managing simulation of mission loads (such as sensors and weapon systems), the five-axis flight turntable is used for simulating attitude changes (such as pitching, rolling, yawing and X/Y/Z axis displacement) of the aircraft, and the moment load simulator is used for simulating dynamic response of the aircraft under pneumatic and mechanical loads.

It should be understood that, in practical application, the semi-physical simulation may also use various integrated hardware systems corresponding to the lightweight nodes and the full-quantization nodes, and the device combination rule may be preset based on the practical use requirement, where if the simulation device is a single hardware device, multiple hardware devices may be combined into a simulation device in multiple hardware system forms according to the device combination rule.

Therefore, the combination form of each simulation model and simulation equipment is diversified and refined, and finally the number-controllable semi-physical simulation nodes are formed, so that on one hand, the actual configuration of the semi-physical simulation nodes in actual application is more fitted, the effectiveness of the semi-physical simulation nodes used in the test is improved, the usability of subsequent performance test data is further improved, on the other hand, complicated and high-precision pairing between the simulation models and the simulation equipment is not needed according to the performance, task scenes and other factors required by the simulation models, the redundant or invalid semi-physical simulation nodes are prevented from generating interference test data, and the complexity of the pairing of the semi-physical simulation nodes, the difficulty and the calculated amount of the subsequent test are effectively reduced.

S103, generating a plurality of cooperation combinations based on the plurality of semi-physical simulation nodes to cooperatively execute a plurality of test tasks to obtain performance test data of each simulation model under different simulation equipment, different cooperation combinations and different task scenes.

Specifically, the collaborative combination refers to a collaborative work unit formed by two or more semi-physical simulation nodes, which is used for simulating the scene of the joint execution task of multiple devices in the real task process, for each test task, a plurality of semi-physical simulation nodes can be selected from multiple semi-physical simulation nodes according to the simulation model requirements of the test task to generate multiple collaborative combinations, the corresponding semi-physical simulation nodes are called to execute for multiple times through different collaborative combinations, the performance of each simulation model under different simulation devices, different collaborative combinations and different task scenes is verified, and the performance test data of each simulation model is recorded. That is, the performance test data is model performance data of the simulation model under specific conditions, for example, the radar model B is respectively loaded into the full-quantization node and the lightweight node, and the detection accuracy of the task scene in the electromagnetic interference or sunny scene, the calculation time consumption of single reasoning and the like.

For example, in unmanned aerial vehicle simulation, the cooperative combination may be a cluster or a team of clusters responsible for performing a specific task, and the cooperative combination may include a first semi-physical simulation node "aircraft model a+full-quantized node", a second semi-physical simulation node "radar model b+lightweight node", and a third semi-physical simulation node "satellite model c+lightweight node".

For example, when the test task is a detection task, the detection task needs to use the capabilities of searching, tracking, time synchronization and the like of the unmanned aerial vehicle, the simulation model of the test task needs to be 3 unmanned aerial vehicle models, a plurality of semi-physical simulation nodes corresponding to the unmanned aerial vehicle models are screened out, and based on the semi-physical simulation nodes, a plurality of collaboration combinations with 3 semi-physical simulation nodes as a group are generated to respectively execute the test task.

In some embodiments, different or the same semi-physical simulation nodes can be combined into a cooperative combination to verify the overall effectiveness of the cooperative work of the isomorphic or heterogeneous resources. That is, the several semi-physical simulation nodes in the collaboration combination may be completely identical, partially identical, and completely different, which is not limited herein.

In some embodiments, the test task or the target simulation task to be executed includes elements such as use requirements, task targets, environment parameters, time constraints, and interaction rules of a plurality of simulation objects. The task type can be determined according to the task targets, and the task type can be predicted by combining partial environment parameters which are public information, such as weather conditions, topography conditions and the like of a task place. Further, according to multiple factors of the task, simulation model requirements can be comprehensively analyzed, and the simulation model requirements can comprise the quantity requirements, the type requirements, the performance requirements and the like of the task on the simulation model.

For example, the usage requirement of the simulation object is 3 unmanned aerial vehicles to cooperatively execute the tracking task of a plurality of target objects, the usage requirement of the simulation model can be determined to be 3 (i.e. the number requirement) unmanned aerial vehicle models (i.e. the type requirement), and the communication delay index of the unmanned aerial vehicle is required to be less than or equal to 50ms (i.e. the performance requirement). For another example, the use requirement of the simulation object is that the tracking task of a plurality of target objects is cooperatively executed in the unmanned aerial vehicle, the use requirement of the simulation model can be determined to be the unmanned aerial vehicle model (i.e. the type requirement), and the communication delay index of the unmanned aerial vehicle is required to be less than or equal to 50ms (i.e. the performance requirement).

The method includes extracting key performance requirements based on task types, task backgrounds and use requirements of simulation objects, for example, detection type tasks need high-precision sensing, cooperative tasks need low-delay communication, mountain scenes need strong anti-interference capability and the like, and setting specific thresholds by combining task execution conditions (such as real-time requirements and resource budget limits) to obtain comprehensive performance requirements. It should be noted that, the task scenario and the simulation model requirement of the test task or the target simulation task to be executed may be determined and modified by the user, or may be obtained by referring to the related art, which is not limited herein.

It should be understood that, for a test task, the environmental parameters of the test task can be stored as task scenes associated with corresponding scene sensitivity indexes, and the simulation model requirements of the test task only consider the quantity requirements and the type requirements of the test task on the simulation model. For the target simulation task, only part of basic information of the environment parameters is the public information, the environment parameters can be gradually disclosed along with the promotion of the simulation task (for example, the environment data obtained after the detection task is executed), the quantity requirements, the type requirements, the performance requirements and the like of the task on the simulation model are required to be comprehensively considered, and the performance requirements can be matched with the simulation model meeting the requirements based on the performance parameter set.

S104, analyzing performance test data of each simulation model, and generating a performance parameter set of each simulation model, wherein the performance parameter set comprises a plurality of scene sensitivity indexes and a plurality of static basic indexes.

Specifically, statistical analysis is performed on all collected performance test data, the value change of the same evaluation index under various conditions is compared, the stability of the performance index under different task scenes, simulation equipment types or cooperation combinations is determined, and further, different evaluation indexes are stored according to stability differentiation, so that a performance parameter set comprehensively reflecting performance characteristics of the performance parameter set is generated, and the performance of each simulation model under various running conditions is systematically described.

The static basic index is the inherent and relatively stable performance of the model, and reflects the basic efficiency of the monomer simulation model. For example, the characteristics of the technical aspects such as the working frequency, the signal processing capability, the data processing capability and the like of the simulation model are kept relatively stable under different conditions. The index is less influenced by external environment, fluctuation under different simulation equipment or task scenes and cooperation combination is small, for example, the 'minimum detection distance' of a radar model is stabilized to be about 5 meters in multiple tests.

The scene sensitivity index is dynamic model performance which varies obviously with external environment or cooperation conditions. For example, characteristics of the radar model at the implementation level, such as the range, coverage, detection accuracy, bit error rate, etc., may vary greatly with environmental changes (such as increased electromagnetic interference intensity or increased terrain complexity). The dynamic performance fluctuation exists in the indexes under a specific task scene, simulation equipment or cooperation combination, for example, although the precision of the simulation model D on the simulation equipment X is high, the cooperation with the simulation model E can cause communication delay to increase rapidly due to resource competition, for example, the data packet loss rate of the communication model can increase rapidly from 0.1% to 5% of the normal environment in an electromagnetic interference scene, or the path planning success rate of the navigation model can be reduced in a light node due to insufficient computing resources in a complex terrain.

In some embodiments, the scene sensitivity index comprises a cooperation index used for representing the overall efficiency of multi-node cooperative work, such as data synchronization time delay between nodes, task cooperative execution success rate and resource competition conflict during multi-model concurrent operation, and a scene index used for representing the influence of environment dynamics on performance, such as terrain matching error in mountain scenes, communication error rate in strong electromagnetic interference and re-planning times of path planning algorithm in multi-obstacle environments.

It should be understood that the performance parameter set provides a basic performance framework of the simulation model through the static basic index, reflects how the simulation model responds to the influence of external factors in a complex and changeable actual operation environment through the scene sensitivity index, and finally decouples the static capacity and the dynamic suitability of the simulation model, so that multidimensional decision basis is provided for subsequent scheduling, and resources and requirements can be accurately matched during dynamic scheduling. For example, in the performance parameter set of the visual recognition model, the static basic index comprises an image processing resolution 1080p, the scene sensitivity index comprises a recognition accuracy rate which is reduced from 92% to 78% under the low illumination condition and can be improved to 85% when being linked with the infrared sensor equipment E, so that a quantitative basis for considering stability and flexibility is provided for subsequent resource scheduling.

In some embodiments, the performance parameter set of each simulation model is associated with a type tag of the simulation model, such as a radar model, an aircraft model, an unmanned aerial vehicle model, etc., for subsequent rapid screening of the required type of simulation model, and its performance parameter set, according to the simulation model requirements.

In some embodiments, the performance test data includes a number of evaluation metrics. The evaluation index is used for quantifying the dimension of performance of the simulation model under specific test conditions (such as task scene type, simulation equipment model and collaborative combination configuration), the evaluation index is embodied by specific numerical values of performance test data, and different types of simulation models can use different evaluation indexes. These evaluation indexes include static characteristics inherent to the model, and dynamic characteristics significantly affected by the external environment.

In some embodiments, S104 comprises comparing a plurality of values of the same evaluation index in performance test data of each simulation model, determining value fluctuation, determining corresponding evaluation indexes as scene sensitive indexes when the value fluctuation is larger than a preset fluctuation value, storing different values of the scene sensitive indexes in association with task scenes and/or simulation equipment and/or cooperation combinations, and determining corresponding evaluation indexes as the static basic indexes when the value fluctuation is smaller than the preset fluctuation value.

The value fluctuation refers to the numerical value variation amplitude of the same evaluation index under different task scenes, simulation equipment or cooperation combinations, and can be quantified by standard deviation, range or relative change rate, and is not limited herein. Correspondingly, a preset fluctuation value can be set for judging the stability degree of the evaluation index, and the specific value can be set according to the actual requirement.

Specifically, the types of the evaluation indexes are distinguished by quantitatively analyzing the numerical value change range of the same evaluation index in performance test data in the execution of multiple test tasks, if the value fluctuation of a certain evaluation index is larger than a preset fluctuation value, the performance of the evaluation index is obviously influenced by the external environment, the evaluation index is judged to be a scene sensitive index, different values of the evaluation index are stored in association with specific influencing factors, such as specific task scene types, simulation equipment models and collaborative combination configuration, a dynamic performance mapping relation is formed, for example, the 'bit error rate 3.5% of the simulation model D is associated to' simulation equipment X 'and mountain scene'. If the value fluctuation of a certain evaluation index is smaller than a preset fluctuation value, the evaluation index is judged to be a static basic index, and the average value or the stable interval of the evaluation index is taken as a fixed parameter.

For example, the calculation time consumption of a certain simulation model on three simulation devices is 15ms, 18ms and 16ms respectively, the value fluctuation is 3ms, the upper limit of the preset fluctuation value is 20% of the median, namely, the upper limit of the preset fluctuation value is a static basic index within the range of 3.2ms of the fluctuation of the median of 16ms, the range exceeding 3.2ms is a scene sensitivity index, and the evaluation index of the calculation time consumption of the certain simulation model can be determined to be the static basic index.

For example, a standard deviation is calculated for the performance test data of each evaluation index across scenes, if the standard deviation exceeds a preset fluctuation value, a scene sensitive mark is triggered, associated metadata (such as an simulation device ID and a collaboration combination number) is extracted, and a performance parameter set with a stability description and dynamic adaptation relation is constructed.

The method is characterized by further comprising the steps of analyzing the change trend of a plurality of values of the scene sensitivity index, respectively determining the influence coefficient of the task scene, the simulation equipment and the cooperation combination on the scene sensitivity index, determining the influence factor of the scene sensitivity index from the task scene, the simulation equipment and the cooperation combination according to the influence coefficient, and storing the influence factor in association with the scene sensitivity index.

Specifically, the variation trend refers to a numerical variation rule of the same index under different test conditions (such as task scene type, simulation equipment model and cooperative combination configuration), for example, the data packet loss rate of a communication model linearly rises along with the increase of electromagnetic interference intensity, and the independent influence degree, namely the influence coefficient, of three factors of task scene, simulation equipment and cooperative combination on the index is quantified according to the variation trend. Correspondingly, a preset coefficient can be set for judging the standard that a certain factor has obvious influence on the index, the specific numerical value can be flexibly set according to the actual requirement, and when the influence coefficient of the certain factor is larger than the preset coefficient, the factor is marked as the influence factor and is associated with the scene sensitive index for storage.

Taking the evaluation index of the detection distance of a radar model as an example, when the simulation equipment A is used, the evaluation index is shortened by 30 percent compared with the simulation equipment B on average, the influence coefficient of the simulation equipment type is 0.3, the influence coefficient of the task scene type is 0.2 when the same index is reduced by 20 percent due to the task scene of the mountain terrain, and the influence coefficient of the cooperation combination type is only 0.08 when the same index only fluctuates by 8 percent in different cooperation combinations. If the preset coefficient is 0.1, the 'cooperative combination' is not included into the influence factors, the influence factors of the scene sensitivity index of the 'detection distance' are 'simulation equipment' and 'task scene', and the 'detection distance' of a certain radar model is correspondingly stored in association with the 'simulation equipment' and the 'task scene'.

The identification and storage of the influence factors can be realized through an automatic data analysis tool, the influence factors can be quantified through a correlation analysis algorithm such as regression analysis or sensitivity analysis, and the influence factors of the same index through the control variable can be identified, for example, under the same task scene and the same simulation equipment, the influence of the cooperative combination is analyzed. The specific algorithm is not limited herein.

The static basic index adopts a differentiation strategy according to the data stability, wherein if the index has extremely small value fluctuation and high consistency in cross-scene test, the index is stored as a fixed value, such as 'memory occupancy rate of 20.3%', if the index has slight fluctuation in a controllable range, an extremum range or a confidence interval is calculated through a plurality of values in statistical performance test data, if the calculation time consumption of a path planning model floats between 10ms and 12ms, the index is stored as a range interval, such as 'time consumption interval [10ms and 12ms ]', based on the statistical distribution, and the communication time consumption of a communication model is respectively 15ms, 15.5ms and 14.8ms in three cooperative combinations, and the range interval is defined as [14.5ms and 15.6ms ].

Therefore, the query efficiency of basic performance is simplified, necessary details are reserved through the upper and lower limits of the interval, an elastic matching space is provided for subsequent scheduling, and meanwhile, the problem of false screening caused by excessively strict single fixed value is avoided.

It should be understood that the finally formed parameter set decouples static capability from dynamic suitability, for example, in the parameter set of the "visual recognition model D", the static basic index includes the "image processing resolution 1080p", and the scene sensitivity index is recorded in combination with the test environment context as "recognition accuracy in low illumination condition is reduced from 92% to 78%, and can be improved to 85% when linked with the infrared sensor device E", so as to provide a quantization basis for subsequent resource scheduling.

Specifically, the priorities of all the simulation tasks to be processed are obtained, and the priorities can be defined by a user or can be automatically generated according to the emergency degree, the resource requirement and the deadline of the tasks, for example, the priorities of rescue tasks are higher than the priorities of patrol tasks, and the priorities are specifically expressed as numerical values or grades. And then, screening one or more tasks with highest priority as 'target simulation tasks' according to the priority labels, ensuring that the key tasks with high timeliness or high value obtain resource matching preferentially under the resource competition scene, and avoiding core flow delay caused by disordered task sequences.

Specifically, for a target simulation task to be executed, the task type can be determined according to the task target, and the task scene, such as weather conditions, topography conditions and the like of a task place, is predicted by combining the currently disclosed environment parameters. Further, according to the use requirement of the simulation object, the requirement of the simulation model, such as the performance requirement and the type requirement of the target simulation task on the simulation model, is determined, the corresponding type of simulation model is screened out firstly based on the type requirement, then the known task scene of the target simulation task is taken as a limiting item, the cooperative combination and the simulation equipment are taken as selectable items, the simulation model with the evaluation index meeting the performance requirement is selected, if the simulation model requirement has a definite quantity requirement on a certain type of simulation model, the corresponding quantity of target simulation models and target simulation equipment with the matching relation can be determined according to the quantity requirement of the target simulation task on the simulation model.

In some embodiments, the performance parameter sets for each simulation model are layer-by-layer screened according to simulation model requirements. The method comprises the steps of firstly, rapidly filtering simulation models with substandard performance requirements through static basic indexes, combining task scenes of target simulation tasks aiming at qualified simulation models with the substandard static basic indexes, taking the determined task scenes in the target simulation tasks as limits, flexibly analyzing specific numerical values of scene sensitivity indexes possibly presented by the simulation models under different task scenes and simulation equipment according to task scenes and simulation equipment related to scene sensitivity indexes of the qualified simulation models, selecting simulation models with lower limit values meeting the performance requirements and combining the simulation models with the simulation equipment, namely, qualified simulation nodes, realizing strong binding screening between the simulation models and the simulation equipment, ensuring that the simulation models and the simulation equipment meet static performance references when in cooperation, and adapting to dynamic environment requirements.

Therefore, the simulation model and the simulation equipment are regarded as an integral execution unit to carry out joint evaluation, the problem that performance does not reach standards due to decoupling of hardware and algorithm in traditional scheduling is fundamentally solved, and performance degradation combination is effectively avoided. For example, although a static index of a simulation model meets the standard, if the simulation model is combined with low-performance equipment, real-time performance is crashed, binding screening can directly exclude the combination, invalid cooperative combination is avoided from being generated later, and the calculation amount and efficiency of resource scheduling are reduced.

Further, based on screening out a plurality of qualified simulation nodes meeting performance requirements, generating a plurality of qualified cooperation combinations, and based on cooperation combinations associated with scene sensitivity indexes of simulation models, evaluating the cooperation among a plurality of simulation models, for example, better or worse performance can be achieved when some combinations cooperate, and at the moment, the simulation model and simulation equipment in the qualified cooperation combinations with the best cooperation can be preferentially selected to serve as a target simulation model and a target simulation equipment.

Therefore, firstly, the simulation model with the unqualified static basic index is eliminated, then the simulation model and the simulation equipment are secondarily screened as the binding execution unit to eliminate the simulation nodes with the unqualified scene sensitive index, finally, the cooperation combination is introduced to carry out the screening of the optimal cooperation efficiency, and the redundancy calculated amount caused by the unqualified low-efficiency simulation nodes and the cooperation combination is effectively reduced through the layer-by-layer screening, so that the efficiency and the response speed of resource scheduling are improved. It should be noted that, for different collaboration combinations, small-amplitude downward fluctuation may occur in the performance of the simulation model due to resource competition and other reasons, but in practical application, more situations are that the collaboration generates better model efficiency, so that even if the collaboration combination is screened and postponed, the selection of the downward fluctuation collaboration combination can be avoided based on the principle of preferential selection, and the finally selected target simulation model and target simulation equipment can still meet the performance requirement.

It is understood that the performance parameter set of each simulation model establishes an index system of dynamic and static separation, and breaks through the resource description mode of 'one-tool cutting'. The static basic index is independent of environmental change and used for reflecting the inherent performance of the model, and the model which does not meet the basic capability requirement is rapidly filtered through the static index in the initial stage of task matching, so that the calculation complexity of the subsequent dynamic evaluation is greatly reduced. And the scene sensitivity index quantifies the performance boundary of the model in real deployment by associating with influencing factors such as task scenes, equipment models and the like, and in the resource binding stage, the optimal model-equipment combination of the scene sensitivity index is preferentially selected in combination with the current task scene. Therefore, the efficiency advantage of traditional resource scheduling is reserved, and the dynamic response capability to complex environments is improved.

In some embodiments, since task scenarios may have diverse parameters, while target simulation tasks may generally disclose or predict some, but not all, task scenarios, scenario-defining parameters (e.g., weather conditions, temperature conditions) may be determined for known task scenarios, and scenario-selectable parameters may be determined for unknown task scenarios. In the subsequent model selection process, the scene sensitivity index can be pre-screened according to the parameter value of the scene limit parameter, the scene limit parameter stored by the scene sensitivity index in an associated mode is matched with the scene limit parameter of the target simulation task (for example, the weather condition of the target simulation task is cloudy days, the weather condition bound by a certain scene sensitivity index is cloudy days), the scene sensitivity index is reserved, the scene limit parameter stored by the scene sensitivity index in an associated mode is not matched with the scene limit parameter of the target simulation task (for example, the limit parameter of the target simulation task is 20 ℃, the weather condition bound by a certain scene sensitivity index is 30-40 ℃) and is set to be in a hidden state, therefore, the known task scene of the target simulation task is taken as a limit item, the performance parameter which is not matched with the task scene of the target simulation task is not used in the resource allocation process of the target simulation task, and the resource allocation efficiency is further improved.

Specifically, the performance parameter set of each simulation model is subjected to multi-dimensional screening according to the simulation model requirements. Firstly, a plurality of unused simulation models in a semi-physical simulation resource library are rapidly filtered through static basic indexes, and the simulation models which do not reach standards are removed. And generating a plurality of alternative simulation nodes based on the plurality of alternative simulation models and a plurality of unused alternative simulation devices in the semi-physical simulation resource library aiming at a plurality of alternative simulation models with the static basic indexes reaching standards. Based on the method, a plurality of alternative simulation nodes are randomly combined to obtain a plurality of alternative cooperation combinations, according to cooperation combinations, task scenes and simulation equipment related to scene sensitivity indexes of the alternative simulation models, specific values of scene sensitivity indexes possibly presented by the simulation models under different alternative cooperation combinations and alternative simulation equipment are flexibly analyzed in combination with task scenes of target simulation tasks, and further comprehensive performance, namely prediction performance, which can be generated by cooperation of the plurality of simulation models is estimated from a macroscopic angle of the whole cooperation combination. The predicted performance refers to the comprehensive performance level expected to be achieved by the alternative collaboration combination in the target simulation task scene, such as task completion time, target recognition accuracy, collaboration time delay and the like.

Meanwhile, the resource occupancy rate can be comprehensively evaluated from the number of simulation models occupied by the alternative collaboration combination, the number of simulation devices, the hardware resource consumption (such as CPU occupation, memory bandwidth and communication bandwidth) required by running simulation and other dimensions, and the predicted performance and the resource occupancy rate are converted into comprehensive scores through a weighting algorithm, for example, the target collaboration combination with the performance reaching the standard and the hardware resource consumption minimum or the performance reaching the standard and the simulation model with the minimum number is preferably selected. Therefore, on one hand, the actual occupancy rate can be accurately calculated, resource overload is avoided, on the other hand, the simulation tasks can be selectively completed by fewer simulation resources, the saved simulation resources can be used for other simulation tasks, and the resource utilization efficiency is maximized.

For example, for an alternative collaborative combination comprising a simulation model F+a simulation device X and a simulation model G+a simulation device Y, by searching that the scene sensitivity index of the simulation model F is 'the strong electromagnetic interference scene recognition rate on the simulation device X is 85%', the scene sensitivity index of the simulation model G is 'the mountain path planning time for 18ms on the simulation device Y', the static basic indexes of the simulation model F and the simulation model G are combined to be 'the data synchronization interval (5 ms)', the comprehensive recognition rate of the alternative collaborative combination in the mountain electromagnetic countermeasure scene of the target simulation task is 82% and the task period is 23ms is predicted by adopting a prediction model.

The resource occupancy rate can be calculated by accumulating static basic indexes of each simulation model in the alternative collaboration combination and resource overhead of simulation equipment, such as 1.2GB of memory occupancy rate of simulation equipment X, 40% of CPU utilization rate of simulation equipment Y and 15% of cross-equipment communication bandwidth occupancy rate, and comprehensively determining the resource occupancy rate by combining the number of simulation models and the number of simulation equipment required to be used by the alternative collaboration combination.

If it is finally determined that the predicted performance of two candidate cooperation combinations meets the performance requirement, wherein the first candidate cooperation combination needs to use 3 simulation models, and the second candidate cooperation combination needs to use only 2 simulation models, and the second candidate cooperation can be selected as the target cooperation combination.

In some embodiments, the predictive model is trained based on a set of marker data of the performance exhibited by the actual task in a large number of collaborative combinations. Thus, the prediction performance of the candidate collaboration combination can be predicted based on the static basic index and the scene sensitivity index of each candidate simulation model in the candidate collaboration combination by the prediction model.

It should be understood that if in the execution process of the simulation task, since a large amount of resources are occupied, the performance parameter set of each simulation model is screened layer by layer according to the requirement of the simulation model (i.e. pre-screening is performed based on static basic indexes, one-round screening is performed based on simulation equipment and task scenes associated with scene sensitive indexes of the simulation model, two-round screening is performed based on cooperative combination and task scenes associated with scene sensitive indexes of the simulation model), so that the simulation model matching the requirement of the performance cannot be selected, and overall screening (i.e. pre-screening is performed based on static basic indexes) can be triggered at this time, and unified screening is performed based on the simulation equipment, task scenes and cooperative combination associated with the scene sensitive indexes of the simulation model. Through the comprehensive overall planning model with larger calculation force and the equipment, and the cooperative efficiency of the model and the model, the simulation model and the simulation equipment are realized, the strong binding screening between the simulation model and the simulation model can select the combination of 'monomer performance is insufficient but combination is complementary' when resources are limited, and the overall performance reaches the standard.

In some embodiments, the simulation model requirements include performance requirements, 105 further includes screening a plurality of unused simulation models based on the performance requirements and the static basic indexes of the simulation models to obtain a plurality of qualified simulation models, generating a plurality of qualified simulation nodes based on the plurality of qualified simulation models and the plurality of unused simulation devices in the semi-physical simulation resource library, evaluating the prediction performance of each qualified simulation node based on the scene sensitivity indexes of the qualified simulation models and the task scenes of the target simulation tasks, selecting the plurality of qualified simulation nodes with the prediction performance meeting the performance requirements, generating a plurality of qualified cooperation combinations based on the screening, and selecting the simulation model and the simulation device in the qualified cooperation combination with the best cooperation under the task scenes of the target simulation tasks as the target simulation model and the target simulation device.

If the qualified simulation nodes with the predicted performance meeting the performance requirements are not selected, for example, an unmanned aerial vehicle model and a visual recognition model are needed at the moment, only the qualified simulation nodes corresponding to the unmanned aerial vehicle model meeting the performance requirements are selected, and the qualified simulation nodes corresponding to the visual recognition model meeting the performance requirements are not selected.

At the moment, a plurality of unused simulation models are screened based on the performance requirements and the static basic indexes of the simulation models to obtain a plurality of alternative simulation models, a plurality of alternative simulation nodes are generated based on the plurality of alternative simulation models and the plurality of unused alternative simulation devices in the semi-physical simulation resource library, a plurality of alternative collaboration combinations are obtained based on the plurality of alternative simulation node combinations, the prediction performance of each alternative collaboration combination in a task scene of a target simulation task is evaluated based on the scene sensitivity indexes of the alternative simulation models, the alternative collaboration combination with the prediction performance meeting the performance requirements is selected to be used as a target collaboration combination, and the alternative simulation models and the alternative simulation devices in the target collaboration combination are used as the target simulation model and the target simulation device. At this time, the target cooperation combination may include a qualified simulation node corresponding to the visual recognition model which does not meet the performance requirement in the layer-by-layer screening, and a simulation node corresponding to another infrared sensor device, where the two simulation nodes cooperate to meet the performance requirement.

It should be noted that, based on the performance requirement and the static basic index of the simulation model, the unused simulation models are screened, and the obtained alternative simulation model or qualified simulation model is essentially the same simulation model, so as to make differential naming for distinguishing layer-by-layer screening and overall screening.

The real-time influence of task scenes, equipment characteristics and cooperation modes on the model performance is quantified through scene sensitivity indexes, the resource scheduling is realized by adopting a layer-by-layer screening and overall screening mechanism based on multiple dimensions (simulation performance and resource occupation), performance degradation or complementary effects under a strong coupling environment are dealt with, performance traps are prejudged and avoided, the limitation of single resource performance evaluation is broken through, and the dynamic response capability of the resource scheduling to the complex environment is optimized. For example, when the resources are sufficient, the performance degradation combination caused by the coupling between the devices or the models is avoided preferentially, when the resources are limited, the performance complementation is realized through overall planning, for example, the performance of a certain model monomer is insufficient, but the task requirement can be met after the model monomer is combined with specific devices or cooperation, so that the resource utilization rate is maximized, the resource waste caused by 'information island' in the traditional scheduling is avoided, the dynamic adaptation is realized under a complex scene, and the efficiency and the expansibility of the simulation system are improved.

In some embodiments, the fast switching when resources are tensed is supported by pre-evaluating the predicted performance and the resource occupancy of the candidate cooperative combination, so as to avoid the rigidness of the traditional scheduling caused by the fixed allocation of the resources.

Specifically, in the simulation task execution stage, the target simulation model is loaded onto the corresponding target simulation equipment to form a plurality of target semi-physical simulation nodes, such as deploying the communication model to the selected communication simulation equipment. And then, in the task running process, the task execution quality of each target semi-physical simulation node is monitored in real time and continuously acquired. The task execution quality is used for representing the actual operation performance of the target semi-physical simulation node and comprises key parameters such as data accuracy (such as target identification rate), time delay (such as communication response time), resource utilization rate (such as CPU occupation rate) and the like. For example, in unmanned aerial vehicle formation simulation, if a certain target semi-physical simulation node is responsible for path planning, the task execution quality may be composed of indexes such as 'path deviation less than or equal to 2 meters', 'task completion time less than or equal to 10 seconds'.

When the task execution quality of a node is lower than the preset quality (such as the path deviation exceeds 3 meters or the task is overtime) through monitoring, the actually measured performance data of the target semi-physical simulation node, namely the actually acquired performance parameters in the current running state, is obtained. And then, comparing the actually measured performance data with expected values of the target simulation model in the historical performance test data, and judging that the target semi-physical simulation node is abnormal if the difference exceeds the preset difference. At this time, the simulation data generated by the node is immediately isolated, for example, the data flow output to the main system is cut off or marked as invalid data, so that the abnormal data is prevented from affecting the whole simulation result. It should be understood that a series of high-correlation performance test data with the highest similarity to the task scene type, collaborative combination configuration and simulation equipment model of the target simulation task can be selected, and the actually measured performance data can be compared.

The setting of the preset quality and the difference threshold can be based on statistical distribution of historical test data, for example, the upper limit of a 95% confidence interval is taken as the preset quality, so that the abnormal response can be ensured, erroneous judgment due to short-term fluctuation is avoided, and specific numerical values are not limited.

It should be appreciated that by dynamically comparing the measured data with the expected data, performance degradation of the simulation model and the simulation device due to environmental fluctuations (e.g., hardware overload, bursty interference) during real operation is accurately identified. For example, if the anti-interference capability of a certain radar model in the test meets the standard, but the performance is suddenly reduced due to poor heat dissipation of equipment in actual operation, the difference between actual data and test data triggers isolation, so that the global simulation is prevented from being polluted by an error detection result.

Specifically, in the simulation task execution stage, if the difference between the actually measured performance data and the historical performance test data of the target semi-physical simulation node is found to be not more than the preset difference by monitoring, the combination of the current simulation model and the simulation equipment still has basic performance guarantee. At this time, the preferred simulation device is selected from among the unused alternative simulation devices in the repository further based on the scene sensitivity index of the model. And then, migrating the target simulation model from the original target simulation equipment to the optimal simulation equipment, and updating the target semi-physical simulation node, so that the resource allocation is optimized while the task quality is ensured.

By way of example, according to the scene sensitivity index of the current target simulation model and the task scene and cooperation combination, simulation equipment with a gain effect on the model performance of the target simulation model is determined and used as a preferred simulation model. That is, the associative memory is the basis for fast positioning and eliminating performance fluctuations in resource allocation, and preferably the simulation device is a simulation device with better performance under the simulation environment (i.e. task scenario, collaboration combination) where the current target simulation model is located, and the selection goal is to ensure that the model maintains or improves performance in a dynamic environment.

It should be appreciated that when a task scene mutation (e.g., a change in weather from clear to heavy rain) or equipment failure causes a slight performance fluctuation, excessive intervention is avoided from triggering, and alternative combinations can be quickly switched according to the association of scene sensitivity indicators, such as model-equipment combinations with higher "heavy rain scene recognition rate" are enabled, without the need for full-scale testing. Therefore, the cooperative potential of the simulation model and the simulation equipment is identified by using the scene sensitivity index, so that the completion quality of the simulation task is improved.

The method comprises the steps of unified management of model resources, a model dynamic loading technology, an exception handling mechanism and the like, wherein the model resource pool construction can achieve unified management and expandability of heterogeneous model resources, reusability and utilization rate of a model are improved, the model dynamic loading technology is used for carrying out dynamic loading of a semi-physical model in simulation and enhancing flexible deployment of the semi-physical simulation model, and the exception handling mechanism can achieve fault-tolerant control in the semi-physical simulation process and improve robustness and stability of a semi-physical simulation system.

In some embodiments, in order to meet the needs of diversified simulation tasks, unified formal description and functional encapsulation are required for simulation models and simulation devices in a semi-physical simulation resource library, specific resources are mapped to logic resources, barriers between collaborative simulation node resources are broken, unified management of the semi-physical simulation resources and comprehensive scheduling in a simulation process are realized, and resource scheduling efficiency and system simulation performance are improved.

Referring to fig. 2, fig. 2 is a schematic diagram of a semi-physical simulation resource library package according to an embodiment of the application. As shown in fig. 2, the simulation resources in the semi-physical simulation resource library include simulation models such as an aircraft model, an unmanned plane model, a satellite model, and a radar model, and further include simulation devices such as a full-scale node 1, a full-scale node 2, a full-scale node N, a lightweight node 1, a lightweight node 2, and a lightweight node N. Test tasks under different task scenes are preset according to different task requirements in an actual application scene, for example, space-time synchronous tasks, transmission tasks, data updating tasks and network transmission tasks are included under cooperative task requirements, and search tasks, target indication tasks and tracking tasks are included under detection task requirements. Based on the implementation steps S102 to S104, scene sensitivity indexes such as the action distance, the coverage area, the detection precision and the bit error rate of the implementation layer are obtained, and static basic indexes such as the working frequency, the signal processing capability and the data processing capability of the technical layer are obtained, and the simulation model is subjected to virtualization packaging from the two layers to obtain the performance parameter set.

It is understood that under the highly adaptive semi-physical simulation scene, the dynamic adaptability and coupling effect of the model and the equipment under different scenes and cooperation combination can capture the implicit dependency relationship among heterogeneous resources under the scene of equipment and field across, so that the formation of 'information island' is avoided. Further, real-time influences of task scenes, device characteristics and collaboration modes on performance of the modules are quantified, resource allocation is matched based on multiple dimensions (simulation performance and resource occupation), and performance degradation or complementary effects under a strong coupling environment (for example, a low-performance module achieves function standard through a specific device combination) are dealt with. And when the resources are limited, the utilization rate is improved through optimization combination, or the problems are quickly isolated and the resources are reconstructed when the task execution is abnormal, so that the expansibility and flexibility of the simulation system are limited finally, and the high-efficiency cooperative requirements in a complex scene are difficult to meet.

Referring to fig. 3, fig. 3 is a schematic block diagram of a computer device according to an embodiment of the present application. The computer device may be a terminal device or a server.

The above-described method may be implemented, for example, in the form of a computer program that is executable on a computer device as shown in fig. 3.

As shown in fig. 3, the computer device includes a processor, a memory, and a network interface connected by a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

The non-volatile storage medium may store an operating system and a computer program. The computer program comprises program instructions which, when executed, cause the processor to perform any one of a number of scheduling methods for semi-physical simulation resources.

The processor is used to provide computing and control capabilities to support the operation of the entire computer device.

The internal memory provides an environment for the execution of a computer program in the non-volatile storage medium that, when executed by the processor, causes the processor to perform any one of a number of scheduling methods for the semi-physical simulation resource.

The network interface is used for network communication such as transmitting assigned tasks and the like.

It should be appreciated that the Processor may be a central processing unit (Central Processing Unit, CPU), it may also be other general purpose processors, digital signal processors (DIGITAL SIGNAL Processor, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like. Wherein the general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

Wherein in one embodiment the processor is configured to run a computer program stored in the memory to implement the steps of:

The steps of the scheduling method for semi-physical simulation resources provided in any embodiment of the present application are not described herein.

The embodiment of the application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, the computer program comprises program instructions, and the processor executes the program instructions to realize the steps of the scheduling method of the semi-physical simulation resource of any one of the embodiment of the application.

The computer readable storage medium may be an internal storage unit of the computer device according to the foregoing embodiment, for example, a hard disk or a memory of the computer device. The computer readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (SMART MEDIA CARD, SMC), a Secure Digital (SD) card, a flash memory card (FLASH CARD), or the like, which are provided on the computer device.

While the application has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the application. Therefore, the protection scope of the application is subject to the protection scope of the claims.

Claims

1. The scheduling method of the semi-physical simulation resource is characterized by comprising the following steps:

S103, generating a plurality of cooperation combinations based on the plurality of semi-physical simulation nodes to cooperatively execute a plurality of test tasks to obtain performance test data of each simulation model in different simulation equipment, different cooperation combinations and different task scenes, wherein the performance test data comprises a plurality of evaluation indexes;

S104, analyzing performance test data of each simulation model to generate a performance parameter set of each simulation model, wherein the performance parameter set comprises a plurality of scene sensitivity indexes and a plurality of static basic indexes, S104 comprises the steps of comparing a plurality of values of the same evaluation index in the performance test data of each simulation model, determining value fluctuation, determining corresponding evaluation indexes as the scene sensitivity indexes when the value fluctuation is larger than a preset fluctuation value, and storing different values of the scene sensitivity indexes in association with task scenes and/or simulation equipment and/or cooperation combinations, and determining corresponding evaluation indexes as the static basic indexes when the value fluctuation is smaller than the preset fluctuation value;

2. The method of claim 1, wherein the method further comprises:

Analyzing the change trend of a plurality of values of the scene sensitivity index, and respectively determining the influence coefficients of a task scene, simulation equipment and a collaboration combination on the scene sensitivity index;

and determining an influence factor of the scene sensitivity index from a task scene, simulation equipment and cooperation combination according to the influence coefficient, and storing the influence factor in association with the scene sensitivity index.

3. The method of claim 1, further comprising storing the static base indicator as a fixed value or a range interval, wherein the range interval is determined based on a plurality of values of the static base indicator.

4. The method of claim 1, wherein the simulation model requirements include performance requirements, the S105 further comprising:

screening a plurality of unused simulation models based on the performance requirements and the static basic indexes of the simulation models to obtain a plurality of alternative simulation models;

Generating a plurality of alternative simulation nodes based on a plurality of alternative simulation models and a plurality of unused alternative simulation devices in the semi-physical simulation resource library, and obtaining a plurality of alternative collaboration combinations based on the plurality of alternative simulation nodes;

Based on scene sensitivity indexes of the alternative simulation models, evaluating the prediction performance and the resource occupancy rate of each alternative collaboration combination in a task scene of the target simulation task;

And selecting the candidate cooperation combination with the predicted performance meeting the performance requirement and the lowest resource occupancy rate as a target cooperation combination, and taking the candidate simulation model and the candidate simulation equipment in the target cooperation combination as the target simulation model and the target simulation equipment.

5. The method of claim 1, wherein the step of S105 is preceded by obtaining a priority of the simulation tasks to be executed, and taking at least one simulation task with the highest priority as the target simulation task.

6. The method of claim 1, wherein the method further comprises:

Loading the target simulation model into corresponding target simulation equipment to obtain a plurality of target semi-physical simulation nodes;

Continuously monitoring task execution quality of a plurality of target semi-physical simulation nodes in the process of executing the target simulation task;

when the task execution quality is lower than a preset quality, acquiring actual measurement performance data of the target semi-physical simulation node;

And if the difference between the actually measured performance data and the performance test data of the target semi-physical simulation node is greater than a preset difference, isolating simulation data generated by the target semi-physical simulation node.

7. The method of claim 6, wherein the method further comprises:

If the difference between the actually measured performance data and the performance test data of the target semi-physical simulation node is smaller than the preset difference, selecting a preferred simulation device from a plurality of unused simulation devices in the semi-physical simulation resource library based on a scene sensitivity index of the target simulation model;

And loading the target simulation model into the optimal simulation equipment to replace the target simulation equipment and update the target semi-physical simulation node.

8. A computer device, the device comprising:

a memory for storing a computer program;

A processor for executing the computer program and for implementing the scheduling method of semi-physical simulation resources according to any one of claims 1 to 7 when the computer program is executed.

9. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program, which when executed by a processor causes the processor to implement the scheduling method of semi-physical simulation resources according to any one of claims 1 to 7.