A method, system, and medium for intelligent auxiliary decision-making in logistics support of equipment systems.

Abstract

A method, system, and medium for intelligent auxiliary decision-making in equipment system logistics support include: S100, acquiring logistics support task data and performing semantic parsing to identify and output task type labels; S200, based on the task type labels, searching a military logistics knowledge base and performing data matching to generate a set of task variables; S300, searching an operations research modeling template library according to the task type labels, matching the modeling template corresponding to the set of task variables, and generating a mathematical modeling file; S400, generating operations research optimization code based on the mathematical modeling file, performing executability verification, and outputting executable code; S500, executing the executable code and solving it through an optimization solver; and S600, acquiring and parsing the output results of the optimization solver to generate at least one logistics support scheme. This invention achieves end-to-end automation from task input to scheme output, improving modeling efficiency.

Description

Technical Field

[0001] This invention relates to the field of large model-assisted decision-making technology, and in particular to an intelligent auxiliary decision-making method, system and storage medium for equipment system logistics support. Background Technology

[0002] Existing logistics support decision-making methods for equipment systems typically employ a technical approach that combines operations research optimization algorithms with human experience. The typical steps are as follows: First, domain experts manually analyze the logistics support task requirements, define decision variables and constraints, and construct an objective function to establish a mathematical model. Next, programmers convert the mathematical model into optimization code that can be executed by a solver. After debugging, the code is solved, and finally, the solution is manually interpreted to form a support plan.

[0003] However, the aforementioned existing technologies have significant limitations: for example, the modeling process is highly dependent on the experience of professionals, resulting in long model building cycles, low efficiency, and difficulty in ensuring consistency and optimality; furthermore, the implementation of the model and code heavily relies on manual conversion, which not only easily introduces logical deviations but also makes it difficult for the model to be updated quickly according to dynamic changes in the battlefield environment, mission parameters, or equipment status.

[0004] In summary, existing technologies cannot achieve end-to-end automation from task input to solution output, resulting in low modeling efficiency and failing to meet the urgent needs of modern military logistics support for intelligence and real-time performance. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides an intelligent auxiliary decision-making method for logistics support of equipment systems, comprising the following steps:

[0006] S100. Acquire logistics support task data and perform semantic parsing to identify and output task type labels; the logistics support task data includes at least one of the following: equipment list, material requirements, transportation route, task time window, and tactical priority.

[0007] S200. Based on the task type tags, retrieve the military logistics knowledge base and perform data matching to generate a task variable set;

[0008] S300. Search the operations research modeling template library according to the task type tag, match the modeling template corresponding to the task variable set, and generate a mathematical modeling file;

[0009] S400. Generate operations research optimization code based on the mathematical modeling file, perform executability verification, and output executable code;

[0010] S500: Execute the executable code and solve it using an optimized solver;

[0011] S600. Obtain and analyze the output of the optimization solver to generate at least one logistics support plan.

[0012] S700: In response to changes in the logistics support task data or external environment, update the task type label and re-execute S100-S600; or update the task variable set and mathematical modeling file and re-execute S400-S600.

[0013] Optionally, the S300 includes the following steps:

[0014] S310. Search the operations research modeling template library according to the task type tag to obtain the modeling template corresponding to the task variable set; the operations research modeling template library includes transportation optimization modeling template, material inventory and supply modeling template, equipment maintenance and scheduling modeling template and cross-domain joint transportation modeling template;

[0015] S320. Map the parameters in the task variable set to the abstract variables defined in the modeling template;

[0016] S330. Based on military rule constraints, generate resource capacity constraints, time window constraints, priority constraints, and cross-domain connection constraints.

[0017] S340. Construct an objective function based on the task type label and generate a mathematical modeling file, wherein the mathematical modeling file includes a set of model variable definitions, a set of constraints, an objective function expression, and the task type label.

[0018] Optionally, the resource capacity constraint condition is specifically that the material capacity is less than or equal to the storage capacity; or, the material capacity is less than or equal to the vehicle capacity.

[0019] The time window constraint is specifically a constraint on the task completion time;

[0020] The priority constraint is specifically that the scheduling order of support tasks for high-priority equipment is higher than that of support tasks for low-priority equipment, wherein the priority of the equipment is preset.

[0021] The cross-domain connectivity constraints specifically refer to the node connectivity conditions between air, sea, or land transportation.

[0022] Optionally, the S400 includes the following steps:

[0023] S410. Parse the mathematical modeling file to generate an initial code framework, including the code header, model variable declaration section, constraint definition section, objective function section, and solution call section;

[0024] S420. Calculate the complexity score according to the preset formula, select the corresponding solver based on the complexity score and the preset solver mapping table, and insert interface statements adapted to the solver into the initial code framework.

[0025] S430. The initial code framework is standardized to generate operations optimization code;

[0026] S440. Execute the operations research optimization code in the sandbox environment to verify whether it can call the solver normally. If the verification fails, trigger the iteration mechanism. If the verification passes, output the executable code.

[0027] Optionally, the verification includes at least syntax correctness verification, logical consistency verification, and solution convergence verification;

[0028] When S440 outputs the executable code, it also outputs the corresponding code version and verification report.

[0029] Optionally, the iterative mechanism specifically includes the following steps:

[0030] S441. Obtain error information;

[0031] S442. Associate the error information with the initial code framework and the mathematical modeling file to generate an error log;

[0032] S443. Obtain and, based on the context information of the error information, correct the initial code framework and regenerate the operations optimization code;

[0033] S444. Execute the regenerated operations research optimization code in the sandbox environment to verify whether it can call the solver normally. If the verification fails, repeat S441-S444. If the verification passes, output the executable code.

[0034] Optionally, the logistics support plan may be a transportation scheduling table, an equipment maintenance priority table, a material supply plan, or a cross-domain collaboration plan.

[0035] The S600 includes the following steps:

[0036] S610. Call the executable code and execute it in a preset runtime environment to obtain the output result;

[0037] S620. Analyze the output results to obtain the assignment results of the task variables and the objective function value, and verify the feasibility of the solution; if the solution is not feasible, prompt the constraint conflict information.

[0038] S630. Based on the task variable assignment results and objective function values ​​of the feasible solutions, the output results are converted into logistics support schemes; wherein, if the output results include multiple candidate solutions, each candidate solution is converted into a logistics support candidate scheme.

[0039] S640. If there are two or more candidate solutions for logistics support, then each candidate solution shall be quantitatively evaluated by a weighted scoring function.

[0040] S650. Based on the weighted scoring results, select one logistics support candidate scheme as the logistics support scheme.

[0041] Optionally, after generating at least one logistics support plan, the method further includes visually displaying the logistics support plan or logistics support candidate plan;

[0042] The visualization includes one or more of the following: map display, Gantt chart display, table display, and chart display.

[0043] Corresponding to the aforementioned intelligent auxiliary decision-making method for equipment system logistics support, the present invention provides an intelligent auxiliary decision-making system for equipment system logistics support, comprising:

[0044] The task parsing module is used to acquire logistics support task data and perform semantic parsing, identify and output task type tags; the logistics support task data includes at least one of the following: equipment list, material requirements, transportation route, task time window, and tactical priority.

[0045] The data matching module is used to retrieve data from the military logistics knowledge base based on the task type tags and perform data matching to generate a set of task variables;

[0046] The modeling template matching module is used to search the operations research modeling template library according to the task type tags, match the modeling template corresponding to the task variable set, and generate a mathematical modeling file;

[0047] The code generation and verification module is used to generate operations research optimization code based on the mathematical modeling file, perform executability verification, and output executable code;

[0048] The solver execution module is used to execute the executable code and solve the problem using an optimized solver;

[0049] The logistics support scheme generation module is used to obtain and parse the output results of the optimization solver and generate at least one logistics support scheme.

[0050] The update module is used to respond to changes in the logistics support task data or external environment, update the task type label, and then trigger the data matching module, modeling template matching module, code generation and verification module, solution execution module and logistics support scheme generation module to re-execute the process; or update the task variable set and mathematical modeling file, and then trigger the code generation and verification module, solution execution module and logistics support scheme generation module to re-execute the process.

[0051] In addition, to achieve the above objectives, the present invention also provides a computer-readable storage medium storing an intelligent auxiliary decision-making program for equipment system logistics support, wherein when the intelligent auxiliary decision-making program for equipment system logistics support is executed by a processor, it implements the steps of the intelligent auxiliary decision-making method for equipment system logistics support as described above.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] (1) Obtaining task type tags through semantic parsing reduces the need for manual analysis of task requirements, improving the efficiency and accuracy of task identification; generating task variable sets based on task type tags from the knowledge base avoids manually defining variables and constraints, lowering the modeling threshold and reducing logical deviations; the use of the operations research modeling template library achieves standardization and modularization of modeling, improving modeling efficiency and model consistency; automatically generating operations research optimization code and performing executable verification eliminates the tedious process of model-to-code conversion relying on manual intervention, improving the accuracy and efficiency of code generation; by optimizing the solver and parsing the results to generate a guarantee scheme, an automated closed loop from task input to scheme output is achieved, greatly shortening the decision-making cycle and ensuring the real-time performance and effectiveness of the guarantee scheme. Furthermore, by responding to changes in task data or the external environment, dynamically updating the corresponding content and executing the corresponding optimization process enhances the system's adaptability to dynamic changes.

[0054] (2) The different modeling templates of this invention are designed for specific logistics task types, such as transportation optimization, material inventory and replenishment, equipment maintenance and scheduling, and cross-domain joint transportation, making the modeling more accurate and in line with actual needs; the parameter mapping step ensures the accurate correspondence between task variables and abstract variables in the modeling template, laying the foundation for the accuracy of subsequent modeling; various constraints are generated based on military rule constraints, such as resource capacity, time window, priority and cross-domain connection constraints, ensuring that the modeling process conforms to the actual military operation specifications and improving the feasibility and practicality of the model; the construction of the objective function further clarifies the optimization direction, making the generated mathematical modeling file more complete, providing high-quality input for subsequent optimization solutions, and helping to obtain better logistics support solutions.

[0055] (3) Resource capacity constraints ensure that the allocation of materials is within the storage and transportation capacity limits, avoiding the problem of resource overload; time window constraints strictly control the task completion time, ensuring that the logistics support tasks are completed on time, which meets the time urgency requirements of military operations; priority constraints follow the principle of prioritizing the scheduling of high-priority equipment support tasks, ensuring the timely support of key equipment and improving combat effectiveness; cross-domain connection constraints standardize the node connection between different modes of transportation, ensuring the smooth operation of cross-domain joint transportation and improving the coordination and continuity of overall logistics support.

[0056] (4) By parsing the mathematical modeling file to generate the initial code framework, the automatic conversion from model to code is realized, which reduces the workload and error probability of manual coding; the appropriate solver is selected according to the complexity score and the preset solver mapping table and the interface statement is inserted to ensure good compatibility between the code and the solver and improve the solving efficiency; the initial code framework is standardized to generate standardized operations research optimization code, which is conducive to code maintenance and expansion; the executableness of the code is verified in the sandbox running environment, and code problems are discovered and solved in a timely manner, ensuring the quality and stability of the output executable code and providing a solid foundation for subsequent optimization and solving.

[0057] (5) By outputting the code version and verification report along with the executable code, it is easier to manage the code version and trace the problem, which provides strong support for subsequent code optimization and system upgrade.

[0058] (6) The iteration mechanism effectively improves the accuracy and reliability of code generation, can correct problems in the code in a timely manner, improves the success rate of code generation, and reduces manual intervention.

[0059] (7) The output results are obtained and parsed by calling the executable code to verify the feasibility of the solution. If the solution is not feasible, the constraint conflict information is prompted, which helps to discover and adjust the problems in the model or data in a timely manner. The feasible solution is converted into a specific logistics support plan, such as a transportation scheduling table, an equipment maintenance priority table, etc., making the abstract optimization results more practical and operable. For multiple candidate solutions, the optimal solution is selected by quantifying the evaluation through a weighted scoring function. The performance of different solutions in each dimension is fully considered, and the scientific selection of support plan is realized, which improves the rationality and effectiveness of logistics support decision-making and provides more valuable decision reference for military commanders.

[0060] (8) The visualization function greatly enhances the intuitiveness and comprehensibility of the logistics support plan. Through various visualization methods such as map display, Gantt chart display, tables, and charts, the complex logistics support plan is presented in a concise and clear manner, which makes it easier for military commanders to quickly grasp the key points of the plan and improve decision-making efficiency. Attached Figure Description

[0061] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0062] Figure 1 This is a simplified flowchart of an embodiment of the intelligent auxiliary decision-making method for logistics support of the equipment system of the present invention;

[0063] Figure 2 This is a framework diagram of an embodiment of the intelligent auxiliary decision-making system for logistics support of the equipment system of the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] like Figure 1 As shown, the present invention provides an intelligent auxiliary decision-making method for logistics support of an equipment system, which includes the following steps:

[0066] S100. Acquire logistics support mission data and perform semantic parsing to identify and output mission type labels; logistics support mission data shall include at least one of the following: equipment list, material requirements, transportation route, mission time window, and tactical priority.

[0067] S200: Based on task type tags, retrieve military logistics knowledge base and perform data matching to generate a set of task variables;

[0068] S300. Search the operations research modeling template library according to the task type tag, match the modeling template corresponding to the task variable set, and generate a mathematical modeling file;

[0069] S400: Generate operations research optimization code based on the mathematical modeling file, perform executability verification, and output executable code;

[0070] S500: Execute the executable code and solve the problem using an optimized solver;

[0071] S600: Obtain and analyze the output of the optimization solver to generate at least one logistics support plan;

[0072] S700: In response to changes in logistics support mission data or the external environment, update the mission type label and re-execute S100-S600; or update the mission variable set and mathematical modeling file and re-execute S400-S600.

[0073] Preferably, changes in the external environment refer to various external factors that require updating the set of logistics support task variables or mathematical modeling files of the equipment system, such as changes in transportation conditions, task constraints, resource availability, equipment status, and consumption.

[0074] This invention obtains task type tags through semantic parsing, reducing the need for manual analysis of task requirements and improving the efficiency and accuracy of task identification. It generates a set of task variables based on a knowledge base retrieved from the task type tags, avoiding manual definition of variables and constraints, lowering the modeling threshold and reducing logical bias. The use of an operations research modeling template library achieves standardization and modularization of modeling, improving modeling efficiency and model consistency. It automatically generates operations research optimization code and verifies its executability, eliminating the tedious manual process of model-to-code conversion and improving the accuracy and efficiency of code generation. By optimizing the solver and parsing the results to generate a guarantee scheme, it achieves an automated closed loop from task input to solution output, significantly shortening the decision-making cycle and ensuring the real-time performance and effectiveness of the guarantee scheme. Furthermore, by responding to changes in task data or the external environment, it dynamically updates corresponding content and executes corresponding optimization processes, enhancing the system's adaptability to dynamic changes.

[0075] In this embodiment, S300 includes the following steps:

[0076] S310. Retrieve the operations research modeling template library based on the task type tag to obtain the modeling template corresponding to the task variable set; the task type tag is transportation, maintenance, supply, or joint; the operations research modeling template library has preset modeling templates for multiple typical logistics support problems, including at least transportation optimization modeling templates, material inventory and supply modeling templates, equipment maintenance and scheduling modeling templates, and cross-domain joint transportation modeling templates; cross-domain joint transportation refers to air / land / sea joint transportation.

[0077] S320. Map the parameters in the task variable set to the abstract variables defined in the modeling template;

[0078] S330. Based on military rule constraints, generate resource capacity constraints, time window constraints, priority constraints, and cross-domain connection constraints.

[0079] S340. Construct an objective function based on the task type label and generate a mathematical modeling file (JSON or intermediate DSL format). The mathematical modeling file includes a set of model variable definitions (including a complete set of definitions of variable symbols, variable types, and variable value ranges), a set of constraints, an objective function expression, and task type labels. Preferably, the construction of the objective function based on the task type label in S340 specifically involves automatically selecting or combining objective functions, such as minimizing total transportation time; minimizing material loss; maximizing task completion rate; and, in multi-objective optimization, performing a weighted combination.

[0080] In this embodiment, the abstract variable is a predefined general symbolic variable in the corresponding modeling template, used to represent elements in the fields of transportation routes, material inventory, equipment maintenance cycles, and cross-domain transportation methods;

[0081] To facilitate understanding of this invention, examples of abstract variables corresponding to the four types of modeling templates are explained below:

[0082] For the transportation optimization modeling template, an example of abstract variables is: V represents the set of vehicles (abstractly representing different types of vehicles), and P represents the set of transportation routes. Indicates the load capacity of a single vehicle, The transportation time along the route. Mapping example: The task variable set includes "5 trucks, each with a load capacity of 10 tons", mapped to the abstract variable... , This indicates that the vehicle type is a truck. This indicates that each vehicle has a load capacity of 10.

[0083] For the material inventory and replenishment modeling template, an example of an abstract variable is: let I represent the set of material types (abstractly representing different types of vehicles). Represents the initial inventory of material i (i∈I), Indicates the demand for material i at time t. This represents the supply quantity of resource i (not shown in the mapping example, but this abstract variable is also a typical abstract variable in the resource inventory and supply modeling template). Mapping example: The task variable set includes "500 rounds of ammunition, with an estimated demand of 200 rounds on day 3", mapped to the abstract variable... , This indicates that the type of supplies is ammunition. , This indicates that the initial inventory is 500 units. This indicates that the required quantity of ammunition on the third day is 200 rounds.

[0084] For the equipment maintenance scheduling modeling template, an example of an abstract variable is: let E represent the equipment set. R represents the mean time between failures (MTBF) of equipment e. This indicates the time required to repair equipment e (e∈E). This represents the resources required for equipment maintenance (not shown in the mapping example, but this abstract variable is also a typical abstract variable in the equipment maintenance scheduling modeling template); Mapping example: The task variable set includes "A certain type of UAV has an average failure interval of 20 hours and requires 2 hours for maintenance", which is mapped to the abstract variable... This indicates that the equipment type is a drone. This indicates that the mean time between failures (MTBF) is 20 hours. This indicates that the repair time is 2 hours.

[0085] For the cross-domain combined transport modeling template: Example of abstract variable: Let M represent the transport set (such as a set of transport modes such as road, rail, and sea transport). R represents the capacity limit of transportation mode m (m∈M); R represents the set of transportation paths; N represents the set of nodes (such as origin, destination, transfer station, port, airport, etc.). This represents the transportation time of mode m on path r (r∈R); This represents the connection time (loading / unloading / transfer waiting time) at node n (n∈N) when switching from transportation mode m1 to transportation mode m2. Mapping example: The task variable set includes "Highway single vehicle weight limit 20 tons", mapped to the abstract variable... , The mode of transport is road, with a capacity limit of 20 tons; the task variable set includes "Sea transport route A→B requires 72 hours" mapped to an abstract variable. , This indicates that the mode of transport is sea freight, and the transport time from route A to B is 72 hours.

[0086] This invention provides different modeling templates for specific logistics task types, such as transportation optimization, material inventory and replenishment, equipment maintenance scheduling, and cross-domain joint transportation, making the modeling more accurate and aligned with actual needs. The parameter mapping step ensures the accurate correspondence between task variables and abstract variables in the modeling template, laying the foundation for the accuracy of subsequent modeling. Various constraints are generated based on military rules, such as resource capacity, time windows, priorities, and cross-domain connectivity constraints, ensuring that the modeling process conforms to actual military operational standards and improving the feasibility and practicality of the model. Constructing the objective function further clarifies the optimization direction, making the generated mathematical modeling file more complete and providing high-quality input for subsequent optimization solutions, thus contributing to obtaining better logistics support solutions.

[0087] In this embodiment, the resource capacity constraint is specifically that the material capacity is less than or equal to the storage capacity; or, the material capacity is less than or equal to the vehicle capacity.

[0088] The time window constraint is specifically a constraint on the task completion time;

[0089] The priority constraint is that the scheduling order of support tasks for high-priority equipment is higher than that of support tasks for low-priority equipment. The priority of the equipment is preset.

[0090] The cross-domain connectivity constraints specifically refer to the node connectivity conditions between air, sea, or land transportation.

[0091] It should be noted that each constraint is automatically generated using standardized mathematical expressions (such as linear constraints / integer constraints).

[0092] This invention ensures that material allocation is within storage and transportation capacity limits through resource capacity constraints, avoiding resource overload; time window constraints strictly control task completion time, ensuring timely completion of logistical support tasks and meeting the time urgency requirements of military operations; priority constraints follow the principle of prioritizing the scheduling of high-priority equipment support tasks, ensuring timely support of critical equipment and improving combat effectiveness; cross-domain coordination constraints standardize the node connections between different modes of transportation, ensuring smooth cross-domain joint transportation and improving the overall coordination and continuity of logistical support.

[0093] In this embodiment, S400 includes the following steps:

[0094] S410. Parse the mathematical modeling file and generate the initial code framework, including the code header (importing libraries, such as frompulp import LpProblem, LpVariable, LpMinimize), the model variable declaration section (automatically defining integer / continuous variables based on the task variable set), the constraint definition section (automatically mapping to model+=... statements), the objective function section (corresponding to the objective expression), and the solver call section (calling solvers such as Gurobi, CPLEX, and OR-Tools).

[0095] S420. Calculate the complexity score according to the preset formula, select the corresponding solver based on the complexity score and the preset solver mapping table, and insert interface statements that are compatible with the solver into the initial code framework.

[0096] S430. Standardize the initial code framework to generate operations research optimization code; preferably, the standardization process includes checking model variable naming rules, unifying mathematical symbols, expanding constraint loops, and code indentation and syntax formatting.

[0097] S440. Execute the operations research optimization code in the sandbox environment to verify whether it can call the solver normally. If the verification fails, trigger the iteration mechanism. If the verification passes, output the executable code.

[0098] Preferably, the complexity score is calculated according to a preset formula, and the corresponding solver is selected based on the complexity score and a preset solver mapping table, including the following steps:

[0099] S421. Calculate the influence value of the decision variable on the difficulty of solving the problem according to the first preset formula. The first preset formula is as follows:

[0100] Where DV represents the total number of decision variables, D Represents the total number of decision variables, preferably ;

[0101] S422. Calculate the influence value of the constraints on the solution difficulty according to the second preset formula. The second preset formula is as follows:

[0102] Where C represents the total number of constraints, Represents the total amount of constraints, preferably ;

[0103] S423. Calculate the influence of integer variables on the difficulty of solving the problem according to the third preset formula. The third preset formula is as follows:

[0104] ,in, Represents the total number of integer variables;

[0105] S424. Calculate the influence of network structure size on problem complexity according to the fourth preset formula. The fourth preset formula is as follows:

[0106] Where NT represents the number of network nodes and A represents the number of network users. Represents the total number of network nodes, preferably , The total number of network users, preferably ;

[0107] S425. The impact of the number of scheduling time steps on the difficulty of solving the problem is calculated according to the fifth preset formula. The fifth preset formula is as follows:

[0108] Where T represents the scheduling time steps, This represents the total number of time steps, preferably 100.

[0109] S426. The complexity score is calculated according to the sixth preset formula, which is as follows:

[0110] ,in, express The corresponding weight is preferably 0.25; express The corresponding weight is preferably 0.25; express The corresponding weight is preferably 0.2; express The corresponding weight is preferably 0.15; express The corresponding weight is preferably 0.15.

[0111] S427. Select the appropriate solver according to the complexity score and the preset solver mapping table.

[0112] For example, the default solver mapping table is shown in Table 1 below:

[0113] Table 1. Example of a preset solver mapping

[0114]

[0115] This invention automates the conversion from model to code by parsing mathematical modeling files to generate an initial code framework, reducing the workload and error probability of manual coding. It selects a suitable solver based on a complexity score and a pre-defined solver mapping table, inserting interface statements to ensure good compatibility between the code and the solver, thus improving solving efficiency. The initial code framework is standardized to generate standardized operations research optimization code, which is beneficial for code maintenance and expansion. The executableness of the code is verified in a sandbox environment, allowing for timely identification and resolution of code problems, ensuring the quality and stability of the output executable code, and providing a solid foundation for subsequent optimization solutions.

[0116] In this embodiment, the verification includes at least syntax correctness verification, logical consistency verification, and solution convergence verification;

[0117] When the S440 outputs executable code, it also outputs the corresponding code version and verification report.

[0118] This invention outputs the code version and verification report along with the executable code, which facilitates code version management and problem tracing, and provides strong support for subsequent code optimization and system upgrades.

[0119] In this embodiment, the iteration mechanism specifically includes the following steps:

[0120] S441. Obtain error information;

[0121] S442. Associate the error messages with the initial code framework and mathematical modeling files to generate an error log;

[0122] S443. Obtain and, based on the context information of the error message, correct the initial code framework and regenerate the operations optimization code;

[0123] S444. Execute the regenerated operations research optimization code in the sandbox environment to verify whether it can call the solver normally. If the verification fails, repeat S441-S444. If the verification passes, output the executable code.

[0124] This invention effectively improves the accuracy and reliability of code generation through an iterative mechanism, can correct problems in the code in a timely manner, increases the success rate of code generation, and reduces manual intervention.

[0125] In this embodiment, the logistics support plan is a transportation scheduling table, an equipment maintenance priority table, a material supply plan, or a cross-domain collaboration plan.

[0126] Preferably, the transportation dispatch table is used to specify the task allocation, departure / arrival time, and route planning for each transportation vehicle;

[0127] The equipment maintenance priority table is used to sort and output maintenance plans according to tactical priority and resource availability.

[0128] The material replenishment plan is used to generate detailed material distribution information, including supply nodes, demand nodes, and replenishment paths;

[0129] Cross-domain collaboration solutions are used to output the connection timing and resource allocation of air, sea, and land transportation;

[0130] The S600 includes the following steps:

[0131] S610. Call the executable code and execute it in the preset runtime environment to obtain the output result; preferably, select the corresponding solver according to the complexity score and the preset solver mapping table. For specific steps, please refer to S421-S427, which will not be elaborated here.

[0132] S620. Parse the output results to obtain the assignment results of task variables (such as vehicle allocation, material allocation quantity, and maintenance task sequence) and objective function values ​​(such as total transportation time and total consumption), and verify the feasibility of the solution (i.e., determine whether there is a solution that satisfies all constraints); if the solution is not feasible, prompt constraint conflict information.

[0133] S630. Based on the assignment results of the task variables of the feasible solution and the objective function value, convert the output result into a logistics support scheme; wherein, if the output result includes multiple candidate solutions, each candidate solution is converted into a logistics support candidate scheme.

[0134] S640. If there are more than two candidate solutions for logistics support, each candidate solution shall be quantitatively evaluated by a weighted scoring function. The weighted scoring function can be preset according to actual needs, and will not be elaborated here.

[0135] S650. Based on the weighted scoring results, select one logistics support candidate scheme as the logistics support scheme.

[0136] This invention obtains and parses the output results by calling executable code, verifying the feasibility of the solution. If the solution is infeasible, it prompts constraint conflict information, which helps to promptly identify and adjust problems in the model or data. It converts feasible solutions into specific logistics support schemes, such as transportation scheduling tables and equipment maintenance priority tables, making the abstract optimization results more practical and operable. For multiple candidate solutions, it uses a weighted scoring function to quantitatively evaluate and select the optimal scheme, fully considering the performance of different schemes in various dimensions, realizing the scientific optimization of support schemes, improving the rationality and effectiveness of logistics support decisions, and providing more valuable decision-making references for military commanders.

[0137] In this embodiment, after generating at least one logistics support plan, the method further includes visually displaying the logistics support plan or a candidate logistics support plan.

[0138] Visualization displays include one or more of the following: map display (showing transportation routes, node status, and time markers on a map), Gantt chart display (visualizing the time schedule of equipment maintenance and transportation tasks), table display, and chart display (generating statistical charts such as task completion rate and material consumption curves).

[0139] This invention significantly enhances the intuitiveness and comprehensibility of logistics support plans through its visualization capabilities. By employing various visualization techniques such as maps, Gantt charts, tables, and graphs, complex logistics support plans are presented in a concise and clear manner, enabling military commanders to quickly grasp the key points and improve decision-making efficiency.

[0140] like Figure 2 As shown, the present invention also provides an intelligent auxiliary decision-making system for equipment system logistics support, which includes:

[0141] The task parsing module 10 is used to acquire logistics support task data and perform semantic parsing, identify and output task type labels; the logistics support task data includes at least one of the following: equipment list, material requirements, transportation route, task time window, and tactical priority.

[0142] Data matching module 20 is used to retrieve military logistics knowledge base and perform data matching based on task type tags to generate a set of task variables;

[0143] The modeling template matching module 30 is used to search the operations research modeling template library according to the task type tags, match the modeling template corresponding to the task variable set, and generate a mathematical modeling file;

[0144] The code generation and verification module 40 is used to generate operations optimization code based on the mathematical modeling file, perform executability verification, and output executable code;

[0145] The solver execution module 50 is used to execute executable code and solve problems through an optimized solver;

[0146] The logistics support scheme generation module 60 is used to obtain and parse the output results of the optimization solver and generate at least one logistics support scheme.

[0147] The update module 70 is used to update the task type label in response to changes in logistics support task data or the external environment, thereby triggering the data matching module 20, modeling template matching module 30, code generation and verification module 40, solution execution module 50 and logistics support scheme generation module 60 to re-execute the process; or to update the task variable set and mathematical modeling file, thereby triggering the code generation and verification module 40, solution execution module 50 and logistics support scheme generation module 60 to re-execute the process.

[0148] This invention also provides a computer-readable storage medium, which may be a computer-readable storage medium included in the memory described in the above embodiments; or it may be a standalone computer-readable storage medium not assembled into a device. The computer-readable storage medium stores at least one instruction, which is loaded and executed by a processor to implement... Figure 1 The illustrated method is an intelligent auxiliary decision-making method for logistics support of the equipment system. The computer-readable storage medium may be a read-only memory, a hard disk, or an optical disk, etc.

[0149] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For system embodiments and storage medium embodiments, since they are basically similar to method embodiments, the descriptions are relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0150] Furthermore, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0151] The foregoing description illustrates and describes preferred embodiments of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the inventive concept by means of the foregoing teachings or techniques or knowledge in related fields. Any modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A method for intelligent auxiliary decision-making in logistics support of an equipment system, characterized in that, Includes the following steps: S100. Acquire logistics support task data and perform semantic parsing to identify and output task type labels; the logistics support task data includes at least one of the following: equipment list, material requirements, transportation route, task time window, and tactical priority. S200. Based on the task type tags, retrieve the military logistics knowledge base and perform data matching to generate a task variable set; S300. Search the operations research modeling template library according to the task type tag, match the modeling template corresponding to the task variable set, and generate a mathematical modeling file; S400. Generate operations research optimization code based on the mathematical modeling file, perform executability verification, and output executable code; S500: Execute the executable code and solve it using an optimized solver; S600. Obtain and analyze the output of the optimization solver to generate at least one logistics support plan. S700: In response to changes in the logistics support task data or external environment, update the task type label and re-execute S100-S600; or update the task variable set and mathematical modeling file and re-execute S400-S600. The S300 includes the following steps: S310. Search the operations research modeling template library according to the task type tag to obtain the modeling template corresponding to the task variable set; the operations research modeling template library includes transportation optimization modeling template, material inventory and supply modeling template, equipment maintenance and scheduling modeling template and cross-domain joint transportation modeling template; S320. Map the parameters in the task variable set to the abstract variables defined in the modeling template; S330. Based on military rule constraints, generate resource capacity constraints, time window constraints, priority constraints, and cross-domain connection constraints. S340. Construct an objective function based on the task type label and generate a mathematical modeling file, wherein the mathematical modeling file includes a set of model variable definitions, a set of constraints, an objective function expression, and the task type label; The resource capacity constraint is specifically that the material capacity is less than or equal to the storage capacity; or, the material capacity is less than or equal to the vehicle capacity. S400 includes the following steps: S410. Parse the mathematical modeling file to generate an initial code framework, including the code header, model variable declaration section, constraint definition section, objective function section, and solution call section; S420. Calculate the complexity score according to the preset formula, select the corresponding solver based on the complexity score and the preset solver mapping table, and insert interface statements adapted to the solver into the initial code framework. S430. The initial code framework is standardized to generate operations optimization code; S440. Execute the operations research optimization code in the sandbox environment to verify whether it can call the solver normally. If the verification fails, trigger the iteration mechanism. If the verification passes, output the executable code. S420 includes the following steps: S421. Calculate the influence value of the decision variable on the difficulty of solving the problem according to the first preset formula. The first preset formula is as follows: Where DV represents the total number of decision variables. Indicates the total number of decision variables; S422. Calculate the influence value of the constraints on the solution difficulty according to the second preset formula. The second preset formula is as follows: Where C represents the total number of constraints, Indicates the total amount of constraints; S423. Calculate the influence of integer variables on the difficulty of solving the problem according to the third preset formula. The third preset formula is as follows: ,in, Represents the total number of integer variables; S424. Calculate the influence of network structure size on problem complexity according to the fourth preset formula. The fourth preset formula is as follows: Where NT represents the number of network nodes and A represents the number of network users. This represents the total number of network nodes. This represents the total number of network users. S425. The influence of the number of scheduling time steps on the difficulty of solving the problem is calculated according to the fifth preset formula. The fifth preset formula is as follows: Where T represents the scheduling time steps, Indicates the total number of time steps; S426. The complexity score is calculated according to the sixth preset formula, which is as follows: ,in, express The corresponding weights; express The corresponding weights; express The corresponding weights; express The corresponding weights; express The corresponding weights; S427. Select the appropriate solver according to the complexity score and the preset solver mapping table; The time window constraint is specifically a constraint on the task completion time; The priority constraint is specifically that the scheduling order of support tasks for high-priority equipment is higher than that of support tasks for low-priority equipment, wherein the priority of the equipment is preset. The cross-domain connectivity constraints are specifically the node connectivity conditions between air, sea, or land transportation. The logistics support plan includes a transportation scheduling table, an equipment maintenance priority table, a material supply plan, or a cross-domain collaboration plan.

2. The intelligent auxiliary decision-making method for equipment system logistics support according to claim 1, characterized in that, The verification includes at least syntax correctness verification, logical consistency verification, and solution convergence verification. When S440 outputs the executable code, it also outputs the corresponding code version and verification report.

3. The intelligent auxiliary decision-making method for equipment system logistics support according to claim 1, characterized in that, The iterative mechanism specifically includes the following steps: S441. Obtain error information; S442. Associate the error information with the initial code framework and the mathematical modeling file to generate an error log; S443. Obtain and, based on the context information of the error information, correct the initial code framework and regenerate the operations optimization code; S444. Execute the regenerated operations research optimization code in the sandbox environment to verify whether it can call the solver normally. If the verification fails, repeat S441-S444. If the verification passes, output the executable code.

4. The intelligent auxiliary decision-making method for equipment system logistics support according to claim 1, characterized in that, The logistics support plan is a transportation scheduling table, an equipment maintenance priority table, a material supply plan, or a cross-domain collaboration plan. The S600 includes the following steps: S610. Call the executable code and execute it in a preset runtime environment to obtain the output result; S620. Analyze the output results to obtain the assignment results of the task variables and the objective function value, and verify the feasibility of the solution; if the solution is not feasible, prompt the constraint conflict information. S630. Based on the task variable assignment results and objective function values ​​of the feasible solutions, the output results are converted into logistics support schemes; wherein, if the output results include multiple candidate solutions, each candidate solution is converted into a logistics support candidate scheme. S640. If there are two or more candidate solutions for logistics support, then each candidate solution shall be quantitatively evaluated by a weighted scoring function. S650. Based on the weighted scoring results, select one logistics support candidate scheme as the logistics support scheme.

5. The intelligent auxiliary decision-making method for equipment system logistics support according to claim 4, characterized in that, After generating at least one logistics support plan, the method also includes visually displaying the logistics support plan or logistics support candidate plan. The visualization includes one or more of the following: map display, Gantt chart display, table display, and chart display.

6. An intelligent auxiliary decision-making system for logistics support of an equipment system, characterized in that, include: The task parsing module is used to acquire logistics support task data and perform semantic parsing, identify and output task type tags; the logistics support task data includes at least one of the following: equipment list, material requirements, transportation route, task time window, and tactical priority. The data matching module is used to retrieve data from the military logistics knowledge base based on the task type tags and perform data matching to generate a set of task variables; The modeling template matching module is used to retrieve the operations research modeling template library according to the task type tag, match the modeling template corresponding to the task variable set, and generate a mathematical modeling file, including: S310, retrieving the operations research modeling template library according to the task type tag to obtain the modeling template corresponding to the task variable set; the operations research modeling template library includes transportation optimization modeling templates, material inventory and supply modeling templates, equipment maintenance and scheduling modeling templates, and cross-domain joint transportation modeling templates; S320, mapping the parameters in the task variable set to the abstract variables defined in the modeling template; S330, generating resource capacity constraints, time window constraints, priority constraints, and other constraints based on military rule constraints. Cross-domain connectivity constraints; S340, construct an objective function based on the task type label and generate a mathematical modeling file, the mathematical modeling file including a set of model variable definitions, a set of constraints, an objective function expression, and the task type label; wherein, the resource capacity constraint is specifically that the material capacity is less than or equal to the storage capacity; or, the material capacity is less than or equal to the vehicle capacity; the time window constraint is specifically a constraint on the task completion time; the priority constraint is specifically that the scheduling order of support tasks for high-priority equipment is higher than that of support tasks for low-priority equipment, wherein the priority of equipment is preset; the cross-domain connectivity constraints are specifically node connectivity conditions between air, sea, or land transportation; The code generation and verification module is used to generate operations research optimization code based on the mathematical modeling file, perform executable verification, and output executable code. The module includes: S410, parsing the mathematical modeling file to generate an initial code framework, including a code header, model variable declaration, constraint definition, objective function, and solver call; S420, calculating a complexity score based on a preset formula, selecting a corresponding solver based on the complexity score and a preset solver mapping table, and inserting interface statements adapted to the solver into the initial code framework; S430, standardizing the initial code framework to generate operations research optimization code; and S440, executing the operations research optimization code in a sandbox environment to verify whether it can correctly call the solver. If the verification fails, an iteration mechanism is triggered; if the verification passes, the executable code is output. S420 includes the following steps: S421. Calculate the influence value of the decision variable on the difficulty of solving the problem according to the first preset formula. The first preset formula is as follows: Where DV represents the total number of decision variables. Indicates the total number of decision variables; S422. Calculate the influence value of the constraints on the solution difficulty according to the second preset formula. The second preset formula is as follows: Where C represents the total number of constraints, Indicates the total amount of constraints; S423. Calculate the influence of integer variables on the difficulty of solving the problem according to the third preset formula. The third preset formula is as follows: ,in, Represents the total number of integer variables; S424. Calculate the influence of network structure size on problem complexity according to the fourth preset formula. The fourth preset formula is as follows: Where NT represents the number of network nodes and A represents the number of network users. This represents the total number of network nodes. This represents the total number of network users. S425. The influence of the number of scheduling time steps on the difficulty of solving the problem is calculated according to the fifth preset formula. The fifth preset formula is as follows: Where T represents the scheduling time steps, Indicates the total number of time steps; S426. The complexity score is calculated according to the sixth preset formula, which is as follows: ,in, express The corresponding weights; express The corresponding weights; express The corresponding weights; express The corresponding weights; express The corresponding weights; S427. Select the appropriate solver according to the complexity score and the preset solver mapping table; The solver execution module is used to execute the executable code and solve the problem using an optimized solver; The logistics support scheme generation module is used to obtain and parse the output results of the optimization solver to generate at least one logistics support scheme; the logistics support scheme is a transportation scheduling table, an equipment maintenance priority table, a material supply scheme, or a cross-domain collaboration scheme. The update module is used to respond to changes in the logistics support task data or external environment, update the task type label, and then trigger the data matching module, modeling template matching module, code generation and verification module, solution execution module and logistics support scheme generation module to re-execute the process; or update the task variable set and mathematical modeling file, and then trigger the code generation and verification module, solution execution module and logistics support scheme generation module to re-execute the process.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an intelligent auxiliary decision-making program for equipment system logistics support, which, when executed by a processor, implements the steps of the intelligent auxiliary decision-making method for equipment system logistics support as described in any one of claims 1 to 5.

Images