A test program set generation method and system

Abstract

The present invention discloses a test program set generation method and system, comprising the following steps: collecting test elements, including test tasks, test resources, and test time; generating a parallel test process under the test resource constraints based on an abstract model of the test task, and then looping and iterating and simulating the parallel test process; generating multiple parallel test processes through multiple optimization algorithms; performing comparative analysis and optimization combination of the multiple parallel test processes to generate an optimal test process; and drawing the optimal test process for all test tasks to generate a parallel test program set. The present invention can implement one or more automatic test program set generation algorithms based on the abstract description of the task and the test resource constraints, thereby reducing the huge workload brought about by various tests of the system, finding the solution that best suits the test system, and optimizing the efficiency of the parallel test process.

Description

Technical Field

[0001] The present invention relates to the technical field of testing equipment, in particular to a test program set generation method and system. Background Art

[0002] From an operational perspective, a missile weapon system is the most fundamental operational unit capable of independently carrying out combat missions. It generally consists of several key components, including missiles, missile launch equipment, target and missile detection and tracking equipment, missile command and control equipment, and missile logistics support equipment. To ensure the performance of the missiles before launch, the reliability and coordination of the missile control system, launch and control circuits, and coordinated signals must be thoroughly inspected and tested before launch. Furthermore, due to the complexity and non-reproducibility of missile weapon systems, testing, simulation, and verification through test systems are crucial to more scientifically and systematically reveal their inherent laws, effectively assessing, predicting, and guiding missile test and launch missions, enhancing the missile's superior operational characteristics, and ensuring the safe and reliable operation of the test system.

[0003] Therefore, whether during the equipment development and reserve phase or during wartime, to optimize the variety, scale, and architecture of missile weapon systems as combat units, it is necessary to fundamentally establish the equipment's tactical and technical performance, reliability, maintainability, supportability, testability, safety, and affordability from the very beginning of its lifecycle. This will optimize its quality and effectiveness, and most importantly, ensure rapid response, testing, and repair times. This urgently requires optimizing resource allocation, shortening design and production cycles, and enhancing intelligent testing and analysis processes.

[0004] Obtaining an efficient and reliable multi-task parallel test system and method suitable for missile weapon systems is not only an urgent requirement of scientific and technological development for this combat unit, but also an important way to improve missile performance and reduce test cycles and resource investment. Driven by this demand, the inventors conducted in-depth research on the characteristics of the system to be tested (the system to be tested is a missile weapon system) and the characteristics of various optimization algorithms. They completed the prototype construction and test verification of a multi-dimensional, multi-task parallel test system, designed a reasonable multi-task, multi-event combined optimization model, and are committed to achieving the optimal test strategy for resource allocation and process scheduling, solving the problem of low test equipment utilization in the original test method (the original test method did not adopt a parallel test process, so the test equipment utilization rate was low).

[0005] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present invention and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to those skilled in the art. Summary of the Invention

[0006] In response to the defects in the prior art, the purpose of the present invention is to provide a test program set generation method and system, which can realize the generation of one or more test program sets based on the abstract description of the task and the test resource constraints, so as to reduce the huge workload brought by various tests of the missile weapon system, find the most suitable solution for the test system, and achieve the optimal efficiency of the multi-dimensional, multi-task parallel testing process, thereby improving the utilization rate of the test equipment and optimizing the efficiency of the parallel testing process.

[0007] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0008] A method for generating a test program set, characterized by comprising the following steps:

[0009] Collect test elements: the test elements include test tasks, test resources and test time;

[0010] Generate an abstract model of the test task according to the test elements for each test task one by one, and perform the following operations to output the optimal test process for each test task;

[0011] Generating a parallel test process under the test resource constraints according to the abstract model of the test task, and then performing cyclic iteration and simulation on the parallel test process;

[0012] Generate multiple parallel test processes through test process planning and combinatorial network model quantitative analysis of multiple optimization algorithms;

[0013] Comparatively analyzing and optimizing the multiple parallel testing processes to generate an optimal testing process;

[0014] Draw the optimal test flow for all test tasks and generate a set of parallel test programs.

[0015] On the basis of the above technical solution, the following steps are also included: during simulation, UML modeling and Petri net modeling are established, and a hierarchical dynamic modeling method of information flow combining UML modeling and Petri net modeling is used and simulation verification is performed;

[0016] Using a directed acyclic graph to describe the execution mode of the test task;

[0017] Generate a directed acyclic graph of parallel test solutions to obtain related solutions, convert them into Petri net expressions for feasibility analysis, and obtain the function value of the parallel test optimization target.

[0018] On the basis of the above technical solutions, the multiple optimization algorithms include: ant colony algorithm, mixed integer programming linear model, large-scale solution technology of meta-heuristic algorithm, and decoding mechanism based on greedy algorithm.

[0019] On the basis of the above technical solutions, it also includes:

[0020] Interface encapsulation steps;

[0021] Encapsulate the platform algorithm interface;

[0022] Store the solution result format file;

[0023] Visualize the scheduling plan.

[0024] Based on the above technical solution, the interface encapsulation step specifically includes:

[0025] A single optimization algorithm and meta-heuristic algorithm are encapsulated as a dynamic link library;

[0026] Packaged into a unified installation file through the distributed architecture of multi-agent algorithm;

[0027] Establish communication connections between multiple Agents by configuring the IP and port of the computer in the distributed architecture;

[0028] The IP and port configurations need to be modified in the pre-defined XML format file.

[0029] Based on the above technical solution, the input interface design step is also included; specifically, it includes:

[0030] Encapsulate the single exact solution and meta-heuristic algorithm in the optimization algorithm platform into a dynamic link library;

[0031] Package files are also compatible with multiple common programming languages;

[0032] The interface function provides the input of basic model data and the modification of parameters related to the performance of the algorithm.

[0033] Based on the above technical solution, the output interface design step is also included; specifically, it includes:

[0034] The dynamic link library dll file output interface provides various indicators of the optimized scheduling plan;

[0035] Includes: the scheduling sequence of each test task on each resource, the scheduling timetable of each test task, and the test end time under the plan;

[0036] Each indicator provides an interface function for data output, which is used for embedded scheduling under different development requirements;

[0037] The optimization algorithm platform provides script file interfaces in various common formats, including XML, as output files for scheduling results;

[0038] Supports automated testing tasks to arrange and start the entire testing process by reading script file data.

[0039] On the basis of the above technical solution, it also includes the steps of visual scheduling display design, including the display of scheduling result data and graphical display;

[0040] The optimization algorithm platform provides various indicator data of scheduling results in the form of tables on the human-computer interaction interface, and supports graphical display of the scheduling arrangements of various test tasks in the form of Gantt charts;

[0041] Among them, Gantt charts and digital tables can be converted into files for subsequent decision-making in the actual test engineering process.

[0042] A test program set generation system, characterized in that the application server in the system includes a storage module and a processing module;

[0043] The test program set generation method is stored in the storage module and executed on the processing module.

[0044] The test program set generation method and system described in the present invention have the following beneficial effects:

[0045] This invention can automatically generate one or more test program sets based on abstract task descriptions and test resource constraints, reducing the enormous workload associated with each test in the system and finding the most suitable solution for the test system, thereby optimizing the efficiency of the parallel testing process. It also implements platform interface encapsulation for the parallel test task scheduling algorithm, supports multi-scenario input interface design, supports multi-scenario output interface design, and visualizes scheduling results. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The present invention has the following accompanying drawings:

[0047] The accompanying drawings are provided for a better understanding of the present invention and are not intended to limit the present invention.

[0048] Figure 1 : Flowchart of the test program set generation method of the present invention;

[0049] Figure 2 : The test program set generation system application structure diagram of the present invention;

[0050] Figure 3 : Parallel test configuration-execution relationship diagram of the test program set generation method of the present invention;

[0051] Figure 4 : Unit test flow chart of the test program set generation system of the present invention;

[0052] Figure 5: The subsystem test flow chart of the test program set generation system of the present invention.

[0053] Figure 6 :Directed acyclic graph of parallel test solutions of the test program set generation method of the present invention

[0054] Figure 7 : Schematic diagram of the multi-agent distributed algorithm framework of the test program set generation method of the present invention. DETAILED DESCRIPTION

[0055] The present invention will be described in further detail below with reference to the accompanying drawings. The detailed description, which is provided for illustrative purposes only and includes various details to aid understanding of the embodiments of the present invention, should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted from the following description.

[0056] like Figure 1 As shown, the present invention provides a test program set generation method, comprising the following steps:

[0057] Collect test elements: the test elements include test tasks, test resources and test time;

[0058] Generate an abstract model of the test task according to the test elements for each test task one by one, and perform the following operations to output the optimal test process for each test task;

[0059] Generating a parallel test process under the test resource constraints according to the abstract model of the test task, and then performing cyclic iteration and simulation on the parallel test process;

[0060] Generate multiple parallel test processes through test process planning and combinatorial network model quantitative analysis of multiple optimization algorithms;

[0061] Comparatively analyzing and optimizing the multiple parallel testing processes to generate an optimal testing process;

[0062] Draw the optimal test flow for all test tasks and generate a set of parallel test programs.

[0063] Among them, the term "test program set" refers to a collection of multiple optimal test processes, test tasks refer to the execution tasks that need to be tested, test resources refer to the components, equipment, parameters and other resources that need to be called in the test tasks, and test time refers to the time required to complete the test.

[0064] The abstract model is generated using a three-element set mapping abstract description method. Specifically, test tasks, test resources, and test time are abstracted into three basic elements. The overall system situation is described through the attributes and quantities of these three basic elements, as well as the mapping relationship between these element sets. A method for expressing the constraints of related test tasks is also established.

[0065] On the basis of the above technical solution, the following steps are also included: during simulation, UML modeling and Petri net modeling are established, and a hierarchical dynamic modeling method of information flow combining UML modeling and Petri net modeling is used and simulation verification is performed;

[0066] Using a directed acyclic graph to describe the execution mode of the test task;

[0067] Generate a directed acyclic graph of parallel test solutions to obtain related solutions, convert them into Petri net expressions for feasibility analysis, and obtain the function value of the parallel test optimization target.

[0068] like Figure 6 As shown, the feasibility analysis involves analyzing the feasibility of test element information. The three basic test elements are test task θ, test resource R, and test time T. A directed acyclic graph (DAG) is used to describe a possible execution method for the task. The circles represent the test task, test resource set, and test time. Furthermore, the arrows in the diagram depict the order in which tasks are transferred. This must satisfy the actual test task constraints, ensuring that a given test resource rj can only be used by one test task at a time. For the scheduling solution, all reachable paths from the start node to the end node are found, and the sum of the test times Ck (k = 1, 2, …, K, assuming there are K reachable paths) corresponding to each reachable path is calculated. max(Ck) is the objective function value corresponding to the solution. The relevant solution can be converted into a Petri net representation for feasibility analysis to obtain the objective function value. This objective function represents the parallel testing optimization goal.

[0069] In order to address the shortcomings of UML in model analysis and verification, we use Petri nets, which have rigorous theoretical analysis methods, to formally verify the model and ensure its accuracy and security. In the test system, Petri nets serve as a bridge between theoretical research and actual system modeling and simulation analysis.

[0070] UML is primarily used in software engineering, providing view models suitable for all stages of development. As a third-generation object modeling language standard, UML integrates the advantages of multiple OOA / OOD methodologies and is the most widely used modeling language standard.

[0071] On the basis of the above technical solutions, the multiple optimization algorithms include: ant colony algorithm, mixed integer programming linear model, large-scale solution technology of meta-heuristic algorithm, and decoding mechanism based on greedy algorithm.

[0072] By comparing and analyzing the results of the above optimization algorithms and optimizing the combination, we can find the final solution suitable for the test system.

[0073] Considering the number of test tasks and test resources, it's often difficult to enumerate and analyze all possible resource scheduling options. By improving the optimization algorithm, the automatically generated test program set is used to plan the test process and path. A computational relationship is established between existing algorithmic techniques and the total completion time for all test tasks, which is then solved as a minimization problem.

[0074] On the basis of the above technical solution, in some optional embodiments, the method further includes:

[0075] Interface encapsulation steps;

[0076] Encapsulate the platform algorithm interface;

[0077] Store the solution result format file;

[0078] Visualize the scheduling plan.

[0079] Based on the above technical solution, the interface encapsulation step specifically includes:

[0080] A single optimization algorithm and meta-heuristic algorithm are encapsulated as a dynamic link library;

[0081] Packaged into a unified installation file through the distributed architecture of multi-agent algorithm;

[0082] Establish communication connections between multiple Agents by configuring the IP and port of the computer in the distributed architecture;

[0083] The IP and port configurations need to be modified in the pre-defined XML format file.

[0084] Based on the above technical solution, the input interface design step is also included; specifically, it includes:

[0085] Encapsulate the single exact solution and meta-heuristic algorithm in the optimization algorithm platform into a dynamic link library;

[0086] Package files are also compatible with multiple common programming languages;

[0087] The interface function provides the input of basic model data and the modification of parameters related to the performance of the algorithm.

[0088] Based on the above technical solution, the output interface design step is also included; specifically, it includes:

[0089] The dynamic link library dll file output interface provides various indicators of the optimized scheduling plan;

[0090] Includes: the scheduling sequence of each test task on each resource, the scheduling timetable of each test task, and the test end time under the plan;

[0091] Each indicator provides an interface function for data output, which is used for embedded scheduling under different development requirements;

[0092] The optimization algorithm platform provides script file interfaces in various common formats, including XML, as output files for scheduling results;

[0093] Supports automated testing tasks to arrange and start the entire testing process by reading script file data.

[0094] On the basis of the above technical solution, it also includes the steps of visual scheduling display design, including the display of scheduling result data and graphical display;

[0095] The optimization algorithm platform provides various indicator data of scheduling results in the form of tables on the human-computer interaction interface, and supports graphical display of the scheduling arrangements of various test tasks in the form of Gantt charts;

[0096] Among them, Gantt charts and digital tables can be converted into documents for subsequent decision-making in the actual test engineering process, thereby reducing the huge workload brought by various tests of missile weapon systems, and at the same time reducing the possibility of erroneous operations caused by human factors.

[0097] The present invention further discloses a test program set generation system using the above method, wherein the application server in the system includes a storage module and a processing module;

[0098] The test program set generation method is stored in the storage module and executed on the processing module.

[0099] The overall functions of the test program set generation system described in the present invention realize the parallel test task scheduling algorithm platform interface encapsulation, input interface design supporting multiple scenarios, output interface design supporting multiple scenarios, and visual display of scheduling results.

[0100] like Figure 2 As shown, in the test program set generation system, the application server includes a storage module and a processing module. The storage module stores a test program set generation method that can be run on the processing module. On this basis, by mining the overall test requirements, data support is provided for the construction of subsequent test process models.

[0101] Further integration Figure 3 As shown, the system includes a task configuration module and a task execution module. The task configuration module generates the test program set and enters the task execution module for execution; parallel test optimization, based on a reasonable division of tasks, fully considers the constraints of test tasks and test resources, minimizes the total completion time of all test tasks, and carries out work from the task configuration level and the task execution level.

[0102] Task configuration level: By studying the automatic generation algorithm of test program sets, a parallel test process under resource constraints is generated according to the abstract model of the task. The process can be iterated and simulated in a loop to achieve the optimal efficiency of the parallel process and output a confirmed parallel test program set.

[0103] Task execution level: Parallel test task execution software calls various test resources based on the imported parallel test program set to achieve reliable and efficient execution of multiple test tasks. Its core is to study the efficient parallel processing and real-time output of multiple tasks under multi-resource reuse.

[0104] In some optional embodiments, combined with Figure 4 As shown, the unit test targets various types of instruments and equipment on the missile. The tests are complex and involve numerous items, including engine ignition safety bolt open and closed function checks, insulation resistance checks, servo mechanism air pressure and oil level tests, and pyrotechnic device circuit and current-limiting resistor checks. To facilitate parallel optimization testing research, typical unit test items were selected for process modeling.

[0105] Further integration Figure 5 As shown in the figure, based on the characteristics of missile test systems, subsystem testing typically includes open-loop and closed-loop control of the process, as well as corresponding static and dynamic testing. This primarily includes testing of the guidance system, attitude control system, power distribution system, and safety self-destruct system, thereby determining the DD test system category.

[0106] like Figure 7 The distributed computing framework based on multi-agent theory can effectively distribute computing tasks to multiple computers.

[0107] The optimization algorithm adopts simulated annealing algorithm SA and evolutionary algorithm EA, and includes the following steps:

[0108] Initialize the resource pool, evolution strategy and calculate the fitness value in the SA algorithm;

[0109] Determine whether to call the migration strategy;

[0110] If the judgment is "yes", individual updates are performed to determine whether all EA algorithms have been executed;

[0111] The EA algorithm runs while the SA algorithm runs;

[0112] The EA algorithm initializes the population, determines the evolution strategy, and calculates the fitness value.

[0113] Determine whether the sending of the best individual and fitness is terminated. If so, enter the SA to determine whether to call the migration strategy step, otherwise repeat the evolution strategy step;

[0114] If the EA has not been fully executed, it will be repeatedly entered to determine whether to call the migration strategy; if the EA is completed, it will end;

[0115] Among them, when entering the individual update, the worst individual among the individuals in SA is adopted to enter the evolutionary strategy step in EA;

[0116] System Pass Figure 7 The algorithm framework diagram is shown in Figure 1. In some embodiments, when scheduling ultra-large-scale test tasks, a slight difference in solution accuracy can affect the total test task duration, potentially resulting in errors measured in days. Therefore, building a distributed computing framework based on multi-agent theory can effectively distribute computing tasks across multiple computers. By fully and effectively utilizing existing idle hardware resources, computing efficiency can be improved while maximizing solution accuracy. Finally, by statistically analyzing various indicators, the priority of algorithm calls at different scales is determined, enabling the computing platform to intelligently match the computing needs of different scales with the algorithms.

[0117] In the meta-heuristic algorithms commonly used in agent theory to construct distributions, the evolutionary strategies among individuals within a population involve the alternating use of multiple operators, aiming to enable the algorithm to quickly converge to a higher-quality solution while avoiding falling into local optima. This section lists classic operators that are highly effective for scheduling problems, focusing on studying the coordination between operators during the evolutionary process to maximize algorithm performance. Furthermore, this section of research also considers the development of new evolutionary strategies, focusing on establishing a diverse library of evolutionary operators to make meta-heuristic algorithms modular and enhance their robustness and flexibility.

[0118] This study investigates a multi-agent distributed framework for rapidly solving ultra-large-scale parallel test task scheduling. For medium- and large-scale task scheduling, metaheuristic-based solution techniques can generally meet efficiency requirements. However, for ultra-large-scale test task scheduling problems, where the total number of available resources and test tasks reaches thousands, these problems pose new challenges to both the efficiency and accuracy of the solution. The performance of a single computer often cannot meet the enormous computational workload, resulting in excessively long solution times. Furthermore, due to the extremely large search domain for feasible solutions to the problem, a single metaheuristic algorithm can easily fall into a local optimum, resulting in significant time consumption and low solution accuracy.

[0119] In addition, for large-scale test task scheduling, a small difference in solution accuracy can affect the total test task time, with the error in time possibly being measured in days. Therefore, building a distributed computing framework based on multi-agent theory can effectively distribute computing tasks to multiple computers. By fully and effectively utilizing existing idle hardware resources, we can improve computing efficiency while maximizing solution accuracy.

[0120] In some optional embodiments, the test program set automatic generation method and system of the present invention can implement the following four functions.

[0121] Parallel Test Task Scheduling Algorithm Platform Interface Encapsulation—This invention primarily focuses on platform algorithm interface encapsulation, solution result format file storage, and scheduling solution visualization. Single precise and metaheuristic algorithms are encapsulated as dynamic link libraries (.dll files), while the multi-agent distributed architecture is packaged as a unified .exe installation file. Communication between multiple agents is established by configuring the IP addresses and ports of the computers in the distributed architecture. These IP and port configurations must be modified in a pre-defined XML format file. The distributed computing system is booted from the host computer where the agent resides.

[0122] Input interface design supports multiple scenarios. To ensure the algorithm can be implemented for embedded development in other engineering projects, the platform's single precise solver and metaheuristic algorithm are encapsulated as a dynamic link library (.dll file). This encapsulated file is also compatible with multiple common programming languages (C, C++, Python, and Java). The interface functions provide input for basic model data and allow modification of parameters that affect algorithm performance.

[0123] Supporting multi-scenario output interface design—The dynamic link library (.dll file) output interface provides various indicators of the optimized scheduling plan. This includes the scheduling sequence of each test task on each resource, the scheduling timetable for each test task, and the test end time under the plan. Each indicator provides an interface function for data output, enabling embedded scheduling based on different development requirements. The algorithm platform provides script file interfaces in various common formats, including XML, as output files for scheduling results. This allows automated test tasks to schedule and initiate the entire test process by reading script file data.

[0124] Visual display of scheduling results—primarily including the display of scheduling result data and graphical presentation. The algorithm platform provides a table display of various scheduling result indicators in the human-computer interface. It also supports the graphical display of the scheduling arrangements of various test tasks in the form of a common Gantt chart. Gantt charts and digital tables can be converted into files for subsequent decision-making in the actual test engineering process.

[0125] The automatic generation of test suites has the following features:

[0126] a) The test efficiency of the generated test program set is optimal;

[0127] b) The convergence time of the algorithm is as short as possible;

[0128] c) It should be possible to obtain the test efficiency value of the algorithm

[0129] The three basic test elements test task θ, test resources R, and test time T are described by a directed acyclic graph and the feasibility analysis is carried out by converting the relevant solutions into Petri net expressions based on this principle. Figure 6 As shown, a feasibility analysis of test element information generates a directed acyclic graph (DAG), describes the test tasks, and expresses them using Petri nets to obtain the objective function value. The ant colony algorithm is used as the basic algorithm for parallel test scheduling optimization. A multi-threaded parallel processing model is constructed to address situations where different test tasks suddenly require the same resource or encounter other conflicts during actual testing. Based on the description of the test assembly, the number of parallel threads to be launched is rationally planned, ensuring reliable synchronization between threads, avoiding frequent resource access switching, and preventing resource deadlock.

[0130] This invention can automatically generate one or more test program sets based on abstract task descriptions and test resource constraints, reducing the enormous workload associated with each test in the system and finding the most suitable solution for the test system, thereby optimizing the efficiency of the parallel testing process. It also implements platform interface encapsulation for the parallel test task scheduling algorithm, supports multi-scenario input interface design, supports multi-scenario output interface design, and visualizes scheduling results.

[0131] The contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0132] The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiment. Any equivalent modifications or changes made by those skilled in the art based on the contents disclosed in the present invention should be included in the protection scope recorded in the claims.

Claims

1. A test program set generation method, characterized in that: The steps include: Collect test elements: the test elements include test tasks, test resources and test time; Generate an abstract model of each test task based on the test elements, and perform the following operations to output the optimal test process for each test task: S1. Generate a parallel test process under the test resource constraints according to the abstract model of the test task, and then perform loop iteration and simulation on the parallel test process: During simulation, UML modeling and Petri net modeling are established, and simulation verification is performed through the information flow hierarchical dynamic modeling method that combines UML modeling and Petri net modeling; Using a directed acyclic graph to describe the execution mode of the test task; Generate a directed acyclic graph of parallel test solutions to obtain relevant solutions, and convert them into Petri net expressions for feasibility analysis; Get the function value of the test parallel test optimization target; S2. Generate multiple parallel test processes through test process planning and combinatorial network model quantitative analysis of multiple optimization algorithms; S3. Comparing and analyzing the multiple parallel test processes and optimizing their combination to generate an optimal test process; S4. Draw the optimal test process for all test tasks and generate a parallel test program set.

2. A test program set generation method according to claim 1, characterized in that: The multiple optimization algorithms include: ant colony algorithm, mixed integer programming linear model, large-scale solution technology of meta-heuristic algorithm, and decoding mechanism based on greedy algorithm.

3. A test program set generation method according to claim 1, characterized in that: Also includes: Interface encapsulation steps; Encapsulate the platform algorithm interface; Store the solution result format file; Visualize the scheduling plan.

4. A test program set generation method according to claim 3, characterized in that: The interface encapsulation step specifically includes: A single optimization algorithm and meta-heuristic algorithm are encapsulated as a dynamic link library; Packaged into a unified installation file through the distributed architecture of multi-agent algorithm; Establish communication connections between multiple Agents by configuring the IP and port of the computer in the distributed architecture; The IP and port configurations need to be modified in the pre-defined XML format file.

5. A test program set generation method according to claim 3, characterized in that: It also includes the input interface design steps; specifically including: Encapsulate the single exact solution and meta-heuristic algorithm in the optimization algorithm platform into a dynamic link library; Package files are also compatible with multiple common programming languages; The interface function provides the input of basic model data and the modification of parameters related to the performance of the algorithm.

6. A test program set generation method according to claim 3, characterized in that: It also includes the output interface design steps; specifically including: The dynamic link library dll file output interface provides various indicators of the optimized scheduling plan; Includes: the scheduling sequence of each test task on each resource, the scheduling timetable of each test task, and the test end time under the plan; Each indicator provides an interface function for data output, which is used for embedded scheduling under different development requirements; The optimization algorithm platform provides script file interfaces in various common formats, including XML, as output files for scheduling results; Supports automated testing tasks to arrange and start the entire testing process by reading script file data.

7. A test program set generation method according to claim 3, characterized in that: It also includes the steps for designing visual scheduling displays, including the display of scheduling result data and graphical presentation; The optimization algorithm platform provides various indicator data of scheduling results in the form of tables on the human-computer interaction interface, and supports graphical display of the scheduling arrangements of various test tasks in the form of Gantt charts; Among them, Gantt charts and digital tables can be converted into files for subsequent decision-making in the actual test engineering process.

8. A test program set generation system, characterized in that: The application server in the system includes a storage module and a processing module; The test program set generation method described in any one of claims 1 to 7 is stored in the storage module and runs on the processing module.

Images