Simulation scheme generation method under uncertain situation based on large model

Abstract

The invention discloses a method for generating a simulation scheme under an uncertain situation based on a large model, which comprises the following steps of: firstly, generating a simulation intention and a guide prompt word which are input into the large model; then, dividing and conquering the simulation scheme from two dimensions of time and resources, and generating an overall framework of the simulation scheme; designing subtask types according to a resource dividing and conquering result; under the overall framework, specifically determining the resource type and the seed task to be executed in each task time period; the method specifically comprises the steps of obtaining a current frame situation; the simulation intention, the guide prompt word, the current situation and sub-tasks generated by all previous frames are input into the large model, and the situation change trend is estimated; inputting a simulation intention, a guide prompt word, a subtask type and a situation change trend estimation result into the large model, and generating a subtask to be executed by the current frame; and finally, integrating the obtained sub-task sequences to be executed by each resource in each task period to generate a simulation scheme.

Description

Technical Field

[0001] This invention relates to the field of simulation scheme generation technology, and in particular to a method for generating simulation schemes under uncertain conditions based on a large model. Background Technology

[0002] Command and control (C2) personnel face a series of complex and dynamic challenges in highly uncertain situations, requiring them to make rapid and informed decisions. The formulation of adversarial strategies is a core element of decision-making; the simulation schemes in this invention primarily refer to adversarial simulation schemes.

[0003] In traditional methods, the generation of such simulation schemes is meticulous and time-consuming, mainly relying on professional knowledge. It has problems such as difficulty in generation and limited game-theoretic capabilities. Existing methods cannot fully solve problems such as multi-step continuous decision-making under uncertain conditions, handling complex scenarios, and task interpretability. Summary of the Invention

[0004] The technical problem solved by this invention is to overcome the shortcomings of the prior art and provide a method for generating simulation schemes under uncertain conditions based on a large model.

[0005] The technical solution of this invention is: Firstly, a method for generating simulation schemes under uncertain conditions based on a large model is provided, comprising:

[0006] Step 1: Generate simulation intent and guiding prompts for inputting into the large model;

[0007] Step 2: Divide and conquer the simulation scheme from the two dimensions of time and resources, complete the division of task time periods and resources, and generate the overall framework of the simulation scheme; and design the sub-task types based on the resource division and conquer results.

[0008] Step 3: Within the overall framework, specifically determine the time period, resources, and sub-tasks to be executed for each task; the specific method is as follows:

[0009] Obtain the current frame situation;

[0010] The simulation intent, guiding prompts, current situation, and sub-tasks generated from all previous frames are input into the large model to estimate the trend of situational changes.

[0011] Then, input the simulation intent, guiding prompts, subtask types, and situation change trend estimation results into the large model to generate the subtask to be executed in the current frame;

[0012] Step 4: Integrate the sequence of subtasks to be executed by each resource in each task time period obtained in Step 3 to generate a simulation scheme.

[0013] Furthermore, in step 2, the simulation scheme is divided and conquered from the time dimension, specifically by dividing it equally according to the total simulation time or by dividing it according to the simulation stage.

[0014] Furthermore, in step 2, the specific method for dividing the simulation scheme from the resource dimension is to divide it by military branch or by organization.

[0015] Furthermore, in step 4, the integrated simulation scheme is saved as a formatted file.

[0016] Furthermore, it also includes step 5, determining whether the number of generated simulation schemes is sufficient. If so, the generated simulation schemes are output; otherwise, steps 2 and 3 are repeated to continue generating the next simulation scheme until the number of simulation schemes meets the requirements.

[0017] Secondly, a simulation scheme generation system based on a large model under uncertain conditions is provided, including:

[0018] The scheme guidance generation module generates simulation intent and guiding prompts for input into the large model;

[0019] The subtask sequence generation module is used to divide and conquer the simulation scheme from both time and resource dimensions, completing the division of task time periods and resources, and generating the overall framework of the simulation scheme; and designing subtask types based on the resource division and consolidation results; under the overall framework, it specifically determines each task time period, the type of resources, and the type of subtask to be executed; the specific method is as follows:

[0020] Obtain the current frame situation;

[0021] The simulation intent, guiding prompts, current situation, and sub-tasks generated from all previous frames are input into the large model to estimate the trend of situational changes.

[0022] Then, input the simulation intent, guiding prompts, subtask types, and situation change trend estimation results into the large model to generate the subtask to be executed in the current frame;

[0023] The interactive scheme editing module is used to integrate the sub-task sequences to be executed by each resource in each task time period obtained by the sub-task sequence generation module to generate a simulation scheme.

[0024] Furthermore, the subtask sequence generation module divides the simulation scheme into two parts based on the time dimension, specifically by dividing it equally according to the total simulation time or by dividing it according to the simulation stages.

[0025] Furthermore, the subtask sequence generation module divides the simulation scheme from a resource perspective by either dividing it by military branch or by organization.

[0026] Furthermore, the interactive scheme editing module saves the integrated simulation scheme as a Word or PDF file.

[0027] Furthermore, the interactive scheme editing module is also used to determine whether the number of generated simulation schemes is sufficient. If so, it outputs the generated simulation schemes; if not, it repeatedly calls the subtask sequence generation module to continue generating the next simulation scheme until the number of simulation schemes meets the requirements.

[0028] The advantages of this invention compared to the prior art are:

[0029] (1) The simulation scheme generation method proposed in this invention solves the problem of action selection in uncertain scenarios. It divides and conquers the simulation scheme from the two dimensions of time and resources to generate the overall framework of the simulation scheme. Then, under the overall framework, it specifically determines the time period of each task, the resources, and the sub-tasks to be executed, and finally forms a sequence of sub-tasks to obtain a complete simulation scheme. Based on the large model, while generating the simulation scheme, it also explains the action selection of each sub-task to ensure that the decision-making process is understandable and verifiable, and facilitates the commander to further screen and modify the scheme.

[0030] (2) This invention proposes a cross-period decision-making action and debugging impact summary feedback method. For each sub-task decision, after the sub-task is generated, the current situation and actions are summarized and added to the scenario prompts when the next round of sub-tasks is generated in the form of prompt words, participating in further situation prediction. In this way, as the situation changes, the sub-task sequence can be given one by one, which can resolve the difficulties caused by the complexity of the scenario.

[0031] (3) The present invention solves the problem of layered generation of simulation schemes in complex scenarios. The simulation schemes generated by the method of the present invention can be directly connected to the simulation simulation platform to complete the game confrontation verification. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the method of the present invention. Detailed Implementation

[0033] To better understand the technical solution of the present invention, the specific embodiments of the present invention are described below.

[0034] This invention proposes a method for generating simulation schemes under uncertain conditions based on a large model, such as... Figure 1 As shown, it includes the following steps:

[0035] Step 1: Generate simulation intent (e.g., whether to adopt an aggressive attack or a conservative defense) and guiding prompts for inputting into the large model;

[0036] Step 2: Divide and conquer the simulation scheme from the two dimensions of time and resources, complete the division of task time periods and resources, and generate the overall framework of the simulation scheme; and design the sub-task types based on the resource division and conquer results.

[0037] The simulation scheme can be divided into two parts based on time, specifically by dividing it into equal parts according to the total simulation time, or by dividing it into simulation phases (for example, the entire simulation scheme includes reconnaissance phase, penetration phase, attack phase, etc., and is divided according to the time required for each phase).

[0038] The specific ways to divide the simulation scheme from the resource dimension are by branch of service (e.g., army, navy, air force) or by organization (e.g., regiment, battalion, company, platoon).

[0039] Based on the results of resource allocation, design sub-task types (for example, by branch of service (aircraft, tank), sub-task types are aerial reconnaissance and ground attack; by organization (first company, second company), sub-task types are westward attack or eastward attack).

[0040] Step 3: Within the overall framework described above, specifically determine the time period, resources, and sub-tasks to be executed for each task (i.e., when, who, and what); the specific method is as follows:

[0041] Obtain the current frame situation;

[0042] The simulation intent, guiding prompts, current situation, and sub-tasks generated from all previous frames are input into the large model to estimate the trend of situational changes. In the initial frame of the simulation, the input does not include sub-tasks generated from all previous frames, but only the simulation intent, guiding prompts, and current situation.

[0043] Then, input the simulation intent, guiding prompts, subtask types, and situation change trend estimation results into the large model to generate the subtask to be executed in the current frame;

[0044] Step 4: Integrate the sequence of subtasks to be executed by each resource in each task time period obtained in Step 3, generate a simulation scheme, and save it in Word or PDF format.

[0045] Step 5: Determine if the number of generated simulation schemes is sufficient. If yes, output the generated simulation schemes. If not, repeat steps 2 and 3 to generate the next simulation scheme until the number of simulation schemes meets the requirements.

[0046] This invention also provides a simulation scheme generation system based on a large model under uncertain conditions, comprising:

[0047] The scheme guidance generation module generates simulation intent and guiding prompts for input into the large model;

[0048] The subtask sequence generation module is used to divide and conquer the simulation scheme from both time and resource dimensions, completing the division of task time periods and resources, and generating the overall framework of the simulation scheme; and designing subtask types based on the resource division and consolidation results; under the overall framework, it specifically determines each task time period, the type of resources, and the type of subtask to be executed; the specific method is as follows:

[0049] Obtain the current frame situation;

[0050] The simulation intent, guiding prompts, current situation, and sub-tasks generated from all previous frames are input into the large model to estimate the trend of situational changes.

[0051] Then, input the simulation intent, guiding prompts, subtask types, and situation change trend estimation results into the large model to generate the subtask to be executed in the current frame;

[0052] The interactive scheme editing module is used to integrate the sub-task sequences to be executed by each resource in each task time period obtained by the sub-task sequence generation module to generate a simulation scheme.

[0053] It is understood that this invention has been described through embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of this invention. Furthermore, under the teachings of this invention, these features and embodiments can be modified to adapt to specific circumstances without departing from the spirit and scope of this invention. Therefore, this invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are protected by this invention.

[0054] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A method for generating simulation schemes under uncertain conditions based on a large model, characterized in that, include: Step 1: Generate simulation intent and guiding prompts for inputting into the large model; Step 2: Divide and conquer the simulation scheme from the two dimensions of time and resources, complete the division of task time periods and resources, and generate the overall framework of the simulation scheme; and design the sub-task types based on the resource division and conquer results. Step 3: Within the overall framework, specifically determine the time period, resources, and sub-tasks to be executed for each task; the specific method is as follows: Obtain the current frame situation; The simulation intent, guiding prompts, current situation, and sub-tasks generated from all previous frames are input into the large model to estimate the trend of situational changes. Then, input the simulation intent, guiding prompts, subtask types, and situation change trend estimation results into the large model to generate the subtask to be executed in the current frame; Step 4: Integrate the sequence of subtasks to be executed by each resource in each task time period obtained in Step 3 to generate a simulation scheme.

2. The method for generating simulation schemes under uncertain conditions based on a large model according to claim 1, characterized in that: In step 2, the simulation scheme is divided and conquered from the time dimension, specifically by dividing it equally according to the total simulation time or by dividing it according to the simulation stage.

3. The method for generating simulation schemes under uncertain conditions based on a large model according to claim 1, characterized in that: In step 2, the simulation scheme is divided into two parts based on resources: either by military branch or by organizational structure.

4. The method for generating simulation schemes under uncertain conditions based on a large model according to claim 1, characterized in that: In step 4, the integrated simulation scheme is saved as a formatted file.

5. The method for generating simulation schemes under uncertain conditions based on a large model according to claim 1, characterized in that: It also includes step 5, which determines whether the number of generated simulation schemes is sufficient. If so, the generated simulation schemes are output; otherwise, steps 2 and 3 are repeated to generate the next simulation scheme until the number of simulation schemes meets the requirements.

6. A simulation scheme generation system for uncertain situations based on a large model, characterized in that, include: The scheme guidance generation module generates simulation intent and guiding prompts for input into the large model; The subtask sequence generation module is used to divide and conquer the simulation scheme from the two dimensions of time and resources, complete the division of task time periods and resources, and generate the overall framework of the simulation scheme. Based on the resource decentralization results, sub-task types are designed; within the overall framework, the specific time period, resources, and sub-tasks to be executed for each task are determined; the specific method is as follows: Obtain the current frame situation; The simulation intent, guiding prompts, current situation, and sub-tasks generated from all previous frames are input into the large model to estimate the trend of situational changes. Then, input the simulation intent, guiding prompts, subtask types, and situation change trend estimation results into the large model to generate the subtask to be executed in the current frame; The interactive scheme editing module is used to integrate the sub-task sequences to be executed by each resource in each task time period obtained by the sub-task sequence generation module to generate a simulation scheme.

7. The simulation scheme generation system for uncertain situations based on a large model according to claim 6, characterized in that: The subtask sequence generation module divides the simulation scheme into two parts based on the time dimension, specifically by dividing it equally according to the total simulation time or by dividing it according to the simulation stages.

8. The simulation scheme generation system for uncertain situations based on a large model according to claim 6, characterized in that: The subtask sequence generation module divides the simulation scheme from a resource perspective, specifically by military branch or by organization.

9. The simulation scheme generation system based on a large model under uncertain conditions according to claim 6, characterized in that: The interactive scheme editing module saves the integrated simulation scheme as a formatted file.

10. The simulation scheme generation system for uncertain situations based on a large model according to claim 6, characterized in that: The interactive scheme editing module is also used to determine whether the number of generated simulation schemes is sufficient. If so, the generated simulation schemes are output; otherwise, the subtask sequence generation module is called repeatedly to continue generating the next simulation scheme until the number of simulation schemes meets the requirements.

Images