A multi-level AI architecture for wargaming simulation

Abstract

This invention discloses a multi-level AI architecture for wargaming, including a computational engine for acquiring battlefield information and transmitting it to a combat intent layer, a task allocation layer, and an operator action layer; receiving and executing actual operator action instructions output by the operator action layer; the battlefield includes the combatant and the enemy; the combat intent layer is used to determine the current battlefield situation based on battlefield information and generate tactical intentions; the task allocation layer is used to generate an operator task list based on battlefield information and tactical intentions; the operator action layer generates actual operator action instructions based on battlefield information and the operator task list; this improves the accuracy of wargaming while reducing the complexity and time consumption of decision-making calculations.

Description

Technical Field

[0001] This invention belongs to the field of wargaming technology, specifically involving a multi-level AI architecture, simulation method, and system for wargaming. Background Technology

[0002] With the development of artificial intelligence technology, how to use artificial intelligence to make reasonable military decisions in complex and ever-changing battlefield environments and adjust them in a timely manner according to changes in the battlefield situation has become a new research direction. The application value and prospects of artificial intelligence technology in the field of military game confrontation are also increasing.

[0003] Based on current research, most games use only one model—behavior tree or finite-state machine—to model the game's agent. Finite-state machines (FSMs) are a common method for modeling the intelligence and behavior of game agents. They abstract complex agent decisions into different states and transitions between them, allowing each state to manage only its own transition conditions, thus reducing the overall complexity of agent decision-making. Behavior trees (BTs) are a mathematical model for planned execution proposed in Next-Gen AI. During execution, behavior trees determine the next execution node by controlling the flow nodes, and make behavioral decisions by executing the behavior nodes.

[0004] However, using finite state machines alone to construct AI can lead to enormous computational demands when faced with complex decision-making tasks. Furthermore, increased model complexity can cause difficulties in implementation and maintenance. Using behavior trees alone for modeling doesn't guarantee that the final choice made by the behavior tree will be optimal, and the result may not be what was intended during modeling; the decision-making process is also time-consuming. Summary of the Invention

[0005] This invention provides a multi-level AI architecture for wargame simulation, which improves the accuracy of wargame simulation while reducing the complexity and time consumption of decision-making calculations.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] The first aspect of this invention provides a multi-level AI architecture for wargaming simulation, including a computational engine, a combat intent layer, a task allocation layer, and an operator action layer:

[0008] The computational engine is used to acquire battlefield information and transmit it to the combat intent layer, task allocation layer, and operator action layer; it receives and executes the actual operator action instructions output by the operator action layer; the battlefield includes the combatant and the enemy.

[0009] The combat intent layer is used to receive battlefield information sent by the temple calculation engine, judge the current battlefield situation based on the battlefield information, and generate tactical intent.

[0010] The task allocation layer is used to receive battlefield information sent by the computational engine and tactical intentions sent by the combat intention layer; and to generate a list of operator tasks based on the battlefield information and tactical intentions.

[0011] The operator action layer is used to receive battlefield information sent by the temple calculation engine and the operator task list sent by the task allocation layer; and to generate actual operator action instructions based on the battlefield information and the operator task list.

[0012] Preferably, the method by which the operational intent layer determines the current battlefield situation and generates tactical intent based on battlefield information includes:

[0013] The battlefield information is used to extract the status of control points, the troop strength of the combat side and the enemy, and the combat score. The battlefield troop strength is the weighted sum of the operators of the combat side or the enemy. The combat score is equal to the sum of the control point score, the remaining operator score, and the combat score of the combat side or the enemy. The control point score is the score obtained by the combat side or the enemy in capturing control points. The remaining operator score is the score of the remaining operators of the combat side or the enemy. The combat score is the score of the combat side or the enemy in destroying the enemy's operators.

[0014] The current battlefield situation is assessed based on the status of control points, the combat strength of both sides, and the combat score, and tactical intentions are generated accordingly; the tactical intentions include offense, all-out offense, defense, and all-out defense.

[0015] Preferably, the method by which the task allocation layer generates the operator task list based on battlefield information and tactical intentions includes:

[0016] Extract the types of operators for each combatant from battlefield information, and determine the effectiveness value for performing various tasks based on each operator type; task types are divided into offensive tasks, defensive tasks, and reconnaissance tasks; obtain the bonus values ​​for different operators to cooperate in completing tasks.

[0017] The operators are randomly combined with offensive, defensive, and reconnaissance tasks to form a combination. Construct a first set of execution tasks; combine the elements in the first set of execution tasks. Combine in pairs to form a combination And calculate the combination Related indicators;

[0018] Based on correlation indicators, the portfolio Perform a redundancy removal operation to obtain a list of operator tasks.

[0019] A second aspect of this invention provides a wargaming simulation method based on a multi-level AI architecture, comprising:

[0020] The combat intent layer is constructed based on finite state machines, the task allocation layer is constructed based on alliance game theory, and the operator action layer is constructed based on behavior trees.

[0021] The system acquires battlefield information and transmits it to the operational intent layer, task allocation layer, and operator action layer; the battlefield includes both the combatant and the enemy.

[0022] Based on battlefield information, the operational intent layer determines the current battlefield situation and generates tactical intent, which is then sent to the task allocation layer. Based on the battlefield information and tactical intent, an operator task list is generated. The operator action layer generates actual operator action instructions based on the battlefield information and the operator task list.

[0023] The actual action instructions of the operators output by the operator action layer are received and executed through the calculation engine.

[0024] Preferably, the method for determining the current battlefield situation and generating tactical intentions based on battlefield information at the operational intent level includes:

[0025] The battlefield information is used to extract the status of control points, the troop strength of the combat side and the enemy, and the combat score. The battlefield troop strength is the weighted sum of the operators of the combat side or the enemy. The combat score is equal to the sum of the control point score, the remaining operator score, and the combat score of the combat side or the enemy. The control point score is the score obtained by the combat side or the enemy in capturing control points. The remaining operator score is the score of the remaining operators of the combat side or the enemy. The combat score is the score of the combat side or the enemy in destroying the enemy's operators.

[0026] The current battlefield situation is assessed based on the status of control points, the combat strength of both sides, and the combat score, and tactical intentions are generated accordingly; the tactical intentions include offense, all-out offense, defense, and all-out defense.

[0027] Preferred methods for determining the current battlefield situation and generating tactical intentions based on the status of battlefield control points, the battlefield strength of the combatant and the enemy, and the combat score include:

[0028] When the combatant's battlefield troop strength reaches N times that of the enemy's battlefield troop strength, it is defined as the combatant having a troop strength advantage; when the enemy's battlefield troop strength reaches N times that of the combatant, it is defined as the combatant having a troop strength disadvantage; otherwise, it is defined as the combatant and the enemy having roughly equal troop strength.

[0029] Will It is defined as the combatant having a score advantage in battle; otherwise, it is defined as the combatant not having a score advantage in battle. It is represented as the sum of the remaining operator scores of the combat side; Expressed as the battle score of the combating side; Expressed as the opponent's score in the battle;

[0030] When the combatant is outnumbered, it adopts a tactical intention of full-scale defense;

[0031] When the combatant has a numerical advantage, if there are non-combatant control points on the battlefield, the offensive tactic is used to capture these points; if the battlefield consists entirely of control points of the combatant, the all-out offensive tactic is used to eliminate the enemy's control points.

[0032] When the combatant and the enemy have roughly equal troop strength, the combatant has a score advantage in the battle, and there is a third party capturing control points, the tactical intention to adopt an offensive approach is to do so.

[0033] When the combatant and the enemy have roughly equal troop strength, the combatant has a score advantage in the battle, and there is no third party to seize control of the point, the tactical intention is to adopt a defensive approach.

[0034] When the combatant and the enemy forces are of equal strength, the combatant does not have an advantage in combat points, and there are non-combatant control points on the battlefield, the tactical intention is to adopt an offensive approach.

[0035] When the combatant and the enemy have roughly equal forces, the combatant does not have an advantage in combat points, and the battlefield is dominated by the combatant's control points, an offensive tactic is adopted.

[0036] Preferred methods for generating a list of operator tasks based on battlefield information and tactical intentions include:

[0037] Extract the types of operators for each combatant from battlefield information, and determine the effectiveness value for performing various tasks based on each operator type; task types are divided into offensive tasks, defensive tasks, and reconnaissance tasks; obtain the bonus values ​​for different operators to cooperate in completing tasks.

[0038] The operators are randomly combined with offensive, defensive, and reconnaissance tasks to form a combination. Construct a first set of execution tasks; combine the elements in the first set of execution tasks. Combine in pairs to form a combination And calculate the combination Related indicators;

[0039] Based on correlation indicators, the portfolio Perform a redundancy removal operation to obtain a list of operator tasks.

[0040] Preferred, calculation combination The formula for expressing the correlation index is:

[0041]

[0042] In the formula, This represents the combination of operators A1 and A2 in the combat side. Related metrics for task execution; This is represented as the additive value of operators A1 and A2 working together to complete the task; This is represented by operator A1, which determines the performance value of the task execution. This is represented by operator A2, which determines the performance value of the task.

[0043] Preferably, the combination is based on the correlation index. Methods for obtaining the operator task list by performing redundancy removal operations include:

[0044] Grouping according to the correlation indicators from high to low Sort the data and retrieve the combinations according to the sort order. Deleting storage involves a combination of storage elements. Other combinations of the inner repetition operator Repeat the redundancy removal operation until all combinations are traversed. Obtain the list of operator tasks.

[0045] Preferably, it further includes: the task allocation layer combining based on correlation indicators. If there are still remaining operators to be assigned after the redundancy removal operation, the remaining operators are assigned tasks according to their performance value.

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0047] This invention determines the current battlefield situation and generates tactical intentions based on battlefield information at the operational intent layer, then sends these intentions to the task allocation layer. It generates an operator task list based on the battlefield information and tactical intentions. The operator action layer then generates actual operator action commands based on the battlefield information and the operator task list. This effectively organizes and coordinates operator actions to ensure the efficient operation of the operator combination, while simultaneously reducing the complexity and time consumption of decision-making calculations.

[0048] In this invention, the operational intent layer is implemented using a finite state machine, responsible for analyzing the battlefield situation, making strategic decisions, and transmitting them to the task allocation layer. The task allocation layer transforms the orders issued by the operational intent layer into more specific instructions, enabling organized offense or defense. This layer uses coalition game theory for decision-making. The operator action layer only needs to execute the specific instructions issued by the task allocation layer. A behavior tree is used to transform these instructions into executable behaviors, allowing for more flexible strategic implementation and improving the accuracy of wargaming simulations. Attached Figure Description

[0049] Figure 1 This is a structural diagram of a multi-level AI architecture for wargaming simulation provided in Example 1;

[0050] Figure 2 This is a structural diagram of a wargaming simulation method based on a multi-level AI architecture provided in Example 2;

[0051] Figure 3 This is a logic diagram of the generation of tactical intent from the operational intent layer provided in Example 2;

[0052] Figure 4 This is a tactical transformation diagram of the wargaming method provided in Example 2. Detailed Implementation

[0053] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0054] Example 1

[0055] like Figure 1 As shown, this embodiment provides a multi-level AI architecture for wargaming simulation, including a strategic calculation engine, a combat intent layer, a task allocation layer, and an operator action layer:

[0056] The computational engine is used to acquire battlefield information and transmit it to the combat intent layer, task allocation layer, and operator action layer; it receives and executes the actual operator action instructions output by the operator action layer; the battlefield includes the combatant and the enemy.

[0057] The operational intent layer is used to receive battlefield information sent by the computational engine, determine the current battlefield situation based on the battlefield information, and generate tactical intents. Specific methods include:

[0058] The battlefield information is used to extract the status of control points, the troop strength of the combat side and the enemy, and the combat score. The battlefield troop strength is the weighted sum of the operators of the combat side or the enemy. The combat score is equal to the sum of the control point score, the remaining operator score, and the combat score of the combat side or the enemy. The control point score is the score obtained by the combat side or the enemy in capturing control points. The remaining operator score is the score of the remaining operators of the combat side or the enemy. The combat score is the score of the combat side or the enemy in destroying the enemy's operators.

[0059] The current battlefield situation is assessed based on the status of control points, the combat strength of both sides, and the combat score, and tactical intentions are generated accordingly. These tactical intentions include offensive, all-out offensive, defensive, and all-out defensive.

[0060] The task allocation layer is used to receive battlefield information sent by the computational engine and tactical intentions sent by the operational intention layer; it generates a list of operator tasks based on the battlefield information and tactical intentions, specifically including:

[0061] Extract the types of operators for each combatant from battlefield information, and determine the effectiveness value for performing various tasks based on each operator type; task types are divided into offensive tasks, defensive tasks, and reconnaissance tasks; obtain the bonus values ​​for different operators to cooperate in completing tasks.

[0062] The operators are randomly combined with offensive, defensive, and reconnaissance tasks to form a combination. Construct a first set of execution tasks; combine the elements in the first set of execution tasks. Combine in pairs to form a combination And calculate the combination Related indicators;

[0063] Based on correlation indicators, the portfolio Perform a redundancy removal operation to obtain a list of operator tasks.

[0064] The operator action layer is used to receive battlefield information sent by the temple calculation engine and the operator task list sent by the task allocation layer; and to generate actual operator action instructions based on the battlefield information and the operator task list.

[0065] Example 2

[0066] like Figures 2 to 3 As shown, this embodiment provides a wargame simulation method based on a multi-level AI architecture. The method in this embodiment can be applied to the architecture described in Embodiment 1. The system includes:

[0067] The combat intent layer is constructed based on finite state machines, the task allocation layer is constructed based on alliance game theory, and the operator action layer is constructed based on behavior trees.

[0068] Acquiring battlefield information and transmitting it to the operational intent layer, task allocation layer, and operator action layer; the methods for acquiring battlefield information including both combatants and adversaries include:

[0069] Methods for determining the current battlefield situation and generating tactical intentions based on battlefield information include:

[0070] The battlefield information is used to extract the status of control points, the troop strength of the combat side and the enemy, and the combat score. The battlefield troop strength is the weighted sum of the operators of the combat side or the enemy. The combat score is equal to the sum of the control point score, the remaining operator score, and the combat score of the combat side or the enemy. The control point score is the score obtained by the combat side or the enemy in capturing control points. The remaining operator score is the score of the remaining operators of the combat side or the enemy. The combat score is the score of the combat side or the enemy in destroying the enemy's operators.

[0071] Based on the status of battlefield control points, the combat strength of both sides, and the combat score, the current battlefield situation is assessed, and tactical intentions are generated. These tactical intentions include offensive, all-out offensive, defensive, and all-out defensive maneuvers. The specific methods are as follows:

[0072] Methods for determining the current battlefield situation and generating tactical intentions based on the status of captured points, the combat strength of both sides, and the combat scores include:

[0073] When the combatant's battlefield troop strength reaches N times that of the enemy's battlefield troop strength, it is defined as the combatant having a troop strength advantage; when the enemy's battlefield troop strength reaches N times that of the combatant, it is defined as the combatant having a troop strength disadvantage; otherwise, it is defined as the combatant and the enemy having roughly equal troop strength.

[0074] Will It is defined as the combatant having a score advantage in battle; otherwise, it is defined as the combatant not having a score advantage in battle. It is represented as the sum of the remaining operator scores of the combat side; Expressed as the battle score of the combating side; Expressed as the opponent's score in the battle;

[0075] When the combatant is outnumbered, it adopts a tactical intention of full-scale defense;

[0076] When the combatant has a numerical advantage, if there are non-combatant control points on the battlefield, the offensive tactic is used to capture these points; if the battlefield consists entirely of control points of the combatant, the all-out offensive tactic is used to eliminate the enemy's control points.

[0077] When the combatant and the enemy have roughly equal troop strength, the combatant has a score advantage in the battle, and there is a third party capturing control points, the tactical intention to adopt an offensive approach is to do so.

[0078] When the combatant and the enemy have roughly equal troop strength, the combatant has a score advantage in the battle, and there is no third party to seize control of the point, the tactical intention is to adopt a defensive approach.

[0079] When the combatant and the enemy forces are of equal strength, the combatant does not have an advantage in combat points, and there are non-combatant control points on the battlefield, the tactical intention is to adopt an offensive approach.

[0080] When the combatant and the enemy have roughly equal forces, the combatant does not have an advantage in combat points, and the battlefield is dominated by the combatant's control points, an offensive tactic is adopted.

[0081] Methods for generating a list of operator tasks based on battlefield information and tactical intentions include:

[0082] Extract the operator types of the combatants from the battlefield information, and determine the effectiveness value for performing various tasks based on each operator type, denoted as... Task types are divided into offensive, defensive, and reconnaissance tasks; the bonus value obtained by different operators cooperating to complete the task is denoted as... ;in, Represented as an operator; Let the task be represented by operator A1; The task is represented by operator A2;

[0083] The operators are randomly combined with offensive, defensive, and reconnaissance tasks to form a combination. Construct a first set of execution tasks; combine the elements in the first set of execution tasks. Combine in pairs to form a combination And calculate the combination The formula for expressing the correlation index is:

[0084]

[0085] In the formula, This represents the combination of operators A1 and A2 installed in the combat side. Related metrics for task execution; This is represented as the additive value of operators A1 and A2 working together to complete the task; This is represented by operator A1, which determines the performance value of the task execution. This is represented by operator A2, which determines the performance value of the task.

[0086] Based on correlation indicators, the portfolio Methods for obtaining the operator task list by performing redundancy removal operations include:

[0087] Grouping according to the correlation indicators from high to low Sort the data and retrieve the combinations according to the sort order. Deleting storage involves a combination of storage elements. Other combinations of the inner repetition operator Repeat the redundancy removal operation until all combinations are traversed. ;

[0088] The task allocation layer combines tasks based on correlation indicators. If there are still remaining operators to be assigned after the redundancy removal operation, assign tasks to the remaining operators according to their performance value; and obtain the operator task list.

[0089] The tactical intent is sent to the task allocation layer; an operator task list is generated based on the battlefield information and the tactical intent; the operator action layer generates actual operator action instructions based on the battlefield information and the operator task list; the actual operator action instructions output by the operator action layer are received and executed through the temple calculation engine.

[0090] The operator action layer is configured with behavioral logic for each operator to autonomously judge the battlefield situation and choose actions according to its own tasks when performing each task. When performing specific tasks, operators need to process and analyze the battlefield data transmitted by the temple calculation engine to obtain information to support action decisions. Only after obtaining information can operators determine their final actions from among many possible actions based on their own tasks.

[0091] like Figure 4 As shown, in this invention, the operational intent layer is implemented using a finite state machine, responsible for analyzing the battlefield situation, making strategic decisions, and transmitting them to the task allocation layer. The task allocation layer transforms the orders issued by the operational intent layer into more specific instructions, enabling organized offense or defense. This layer uses coalition game theory for decision-making. For the operator action layer, it only needs to execute the specific instructions issued by the task allocation layer. A behavior tree is used to transform commands into executable behaviors, allowing for more flexible tactical transformations and strategic implementation, thus improving the accuracy of wargaming simulations.

[0092] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0094] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0095] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0096] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A multi-level AI architecture for wargaming simulation, characterized in that, This includes the strategic engine, operational intent layer, task allocation layer, and operator action layer: The computational engine is used to acquire battlefield information and transmit it to the combat intent layer, task allocation layer, and operator action layer; it receives and executes the actual operator action instructions output by the operator action layer; the battlefield includes the combatant and the enemy. The combat intent layer is used to receive battlefield information sent by the computing engine, judge the current battlefield situation based on the battlefield information, and generate tactical intent. The task allocation layer is used to receive battlefield information sent by the computing engine and tactical intentions sent by the combat intention layer. Methods for generating a list of operator tasks based on battlefield information and tactical intentions include: Extract the types of operators for each combatant from battlefield information, and determine the effectiveness value for performing various tasks based on each operator type; task types are divided into offensive tasks, defensive tasks, and reconnaissance tasks; obtain the bonus values ​​for different operators to cooperate in completing tasks. The operators are randomly combined with offensive, defensive, and reconnaissance tasks to form a combination. Construct a first set of execution tasks; combine the elements in the first set of execution tasks. Combine in pairs to form a combination And calculate the combination The related indicators; the formula is: ； In the formula, This represents the combination of operators A1 and A2 in the combat side. Related metrics for task execution; This is represented as the additive value of operators A1 and A2 working together to complete the task; This is represented by operator A1, which determines the performance value of the task execution. This is represented by operator A2, which determines the performance value of the task execution. Based on correlation indicators, the portfolio Performing a redundancy removal operation yields a list of operator tasks, specifically including: Grouping according to the correlation indicators from high to low Sort the data and retrieve the combinations according to the sort order. Deleting storage involves a combination of storage elements. Other combinations of the inner repetition operator Repeat the redundancy removal operation until all combinations have been traversed. Obtain the list of operator tasks; The operator action layer is used to receive battlefield information sent by the temple calculation engine and the operator task list sent by the task allocation layer; and to generate actual operator action instructions based on the battlefield information and the operator task list.

2. The multi-level AI architecture for wargaming simulation according to claim 1, characterized in that, The method by which the operational intent layer determines the current battlefield situation and generates tactical intent based on battlefield information includes: The battlefield information is used to extract the status of control points, the troop strength of the combat side and the enemy, and the combat score. The battlefield troop strength is the weighted sum of the operators of the combat side or the enemy. The combat score is equal to the sum of the control point score, the remaining operator score, and the combat score of the combat side or the enemy. The control point score is the score obtained by the combat side or the enemy in capturing control points. The remaining operator score is the score of the remaining operators of the combat side or the enemy. The combat score is the score of the combat side or the enemy in destroying the enemy's operators. The current battlefield situation is assessed and tactical intentions are generated based on the status of control points, the combat strength of both sides, and the combat scores; the tactical intentions include offense, all-out offense, defense, and all-out defense.

3. A wargaming simulation method based on a multi-level AI architecture, characterized in that, include: The combat intent layer is constructed based on finite state machines, the task allocation layer is constructed based on alliance game theory, and the operator action layer is constructed based on behavior trees. The computational engine acquires battlefield information and transmits it to the combat intent layer, task allocation layer, and operator action layer; the battlefield includes the combatant and the enemy. Based on battlefield information, the operational intent layer judges the current battlefield situation and generates tactical intent, which is then sent to the task allocation layer. The task allocation layer generates a list of operator tasks based on battlefield information and tactical intentions, specifically including: Extract the types of operators for each combatant from battlefield information, and determine the effectiveness value for performing various tasks based on each operator type; task types are divided into offensive tasks, defensive tasks, and reconnaissance tasks; obtain the bonus values ​​for different operators to cooperate in completing tasks. The operators are randomly combined with offensive, defensive, and reconnaissance tasks to form a combination. Construct a first set of execution tasks; combine the elements in the first set of execution tasks. Combine in pairs to form a combination And calculate the combination The related indicators; the formula is: ； In the formula, This represents the combination of operators A1 and A2 in the combat side. Related metrics for task execution; This is represented as the additive value of operators A1 and A2 working together to complete the task; This is represented by operator A1, which determines the performance value of the task execution. This is represented by operator A2, which determines the performance value of the task execution. Based on correlation indicators, the portfolio Performing a redundancy removal operation yields a list of operator tasks, specifically including: Grouping according to the correlation indicators from high to low Sort the data and retrieve the combinations according to the sort order. Deleting storage involves a combination of storage elements. Other combinations of the inner repetition operator Repeat the redundancy removal operation until all combinations have been traversed. Obtain the list of operator tasks; The operator action layer generates actual operator action instructions based on battlefield information and the operator task list; the temple calculation engine executes the actual operator action instructions output by the operator action layer.

4. The wargaming simulation method based on a multi-level AI architecture according to claim 3, characterized in that, Methods for determining the current battlefield situation and generating tactical intentions based on battlefield information at the operational intent level include: The battlefield information is used to extract the status of control points, the troop strength of the combat side and the enemy, and the combat score. The battlefield troop strength is the weighted sum of the operators of the combat side or the enemy. The combat score is equal to the sum of the control point score, the remaining operator score, and the combat score of the combat side or the enemy. The control point score is the score obtained by the combat side or the enemy in capturing control points. The remaining operator score is the score of the remaining operators of the combat side or the enemy. The combat score is the score of the combat side or the enemy in destroying the enemy's operators. The current battlefield situation is assessed and tactical intentions are generated based on the status of control points, the combat strength of both sides, and the combat scores; the tactical intentions include offense, all-out offense, defense, and all-out defense.

5. The wargaming simulation method based on a multi-level AI architecture according to claim 4, characterized in that, Methods for determining the current battlefield situation and generating tactical intentions based on the status of captured points, the combat strength of both sides, and the combat scores include: When the combatant's battlefield troop strength reaches N times that of the enemy's battlefield troop strength, it is defined as the combatant having a troop strength advantage; when the enemy's battlefield troop strength reaches N times that of the combatant, it is defined as the combatant having a troop strength disadvantage; otherwise, it is defined as the combatant and the enemy having roughly equal troop strength. Will It is defined as the combatant having a score advantage in battle; otherwise, it is defined as the combatant not having a score advantage in battle. It is represented as the sum of the remaining operator scores of the combat side; Expressed as the battle score of the combating side; Expressed as the opponent's score in the battle; When the combatant is outnumbered, it adopts a tactical intention of full-scale defense; When the combatant has a numerical advantage, if there are non-combatant control points on the battlefield, the offensive tactic is used to capture these points; if the battlefield consists entirely of control points of the combatant, the all-out offensive tactic is used to eliminate the enemy's control points. When the combatant and the enemy have roughly equal troop strength, the combatant has a score advantage in the battle, and there is a third party capturing control points, the tactical intention to adopt an offensive approach is to do so. When the combatant and the enemy have roughly equal troop strength, the combatant has a score advantage in the battle, and there is no third party to seize control of the point, the tactical intention is to adopt a defensive approach. When the combatant and the enemy forces are of equal strength, the combatant does not have an advantage in combat points, and there are non-combatant control points on the battlefield, the tactical intention is to adopt an offensive approach. When the combatant and the enemy have roughly equal forces, the combatant does not have an advantage in combat points, and the battlefield is dominated by the combatant's control points, an offensive tactic is adopted.

6. The wargaming simulation method based on a multi-level AI architecture according to claim 3, characterized in that, Also includes: The task allocation layer combines tasks based on correlation indicators. If there are still remaining operators to be assigned after the redundancy removal operation, the remaining operators are assigned tasks according to their performance value.

Images