Intelligent decision-making method and system for cooperative game of hybrid unmanned aerial vehicle formation
By calculating the strict game matrix and obtaining the initial hybrid strategy Nash equilibrium, the problem that the existing technology cannot make decisions on the coordinated work of attack and interference drones is solved, effective game decisions on their coordinated work are realized, and coordinated operations are improved efficiency.
Patent Information
- Application Number
- CN202411005769.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-15
AI Technical Summary
The existing technology cannot make game decisions on the coordinated work of attack drones and jamming drones, which increases the complexity of coordinated operations.
By obtaining the data of the game participants, calculating the strict game matrix, and obtaining the initial mixed strategy Nash equilibrium based on the matrix, selecting the initial strategy with the greatest probability, generating a strategy set, and updating the matrix to obtain the mixed strategy with the greatest profit, realizing attack-type and interference-type Game decision-making of drones working together.
The game decisions for attack and interference drones are realized when working together, and the efficiency and effectiveness of coordinated operations are improved.
Smart Images

Figure CN118963410A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of security technology, and in particular to an intelligent decision-making method and system for hybrid unmanned aerial vehicle formation collaborative game. Background Art
[0002] Drones can autonomously complete given tasks in a variety of environments, effectively reducing labor costs. Compared with manned aircraft, drones have the characteristics of no casualties, strong maneuverability, light weight, and low cost, and have become one of the important members of modern air combat weapons and equipment. Due to the limited number and performance of weapons and sensors carried by a single drone, its ability to perform air combat missions is correspondingly limited, and multiple drones can better complete air combat missions through effective coordination. Therefore, the coordinated execution of tasks by multiple drones has received more and more attention. How to set up a game decision for drone formations to coordinate operations is currently a major problem.
[0003] When setting up a game decision for cooperative combat for a UAV formation, it is generally necessary to assign a target to each UAV. In the prior art, a zero-sum matrix game model is constructed and solved to determine the optimal solution as the decision result. The mainstream solution algorithms include the Lemke-Howson (LH) algorithm, the mixed integer programming (MIP) algorithm, and the Double Oracle (DO) algorithm.
[0004] However, the attacker's drones may include attack drones and jammer drones. Attack drones attack high-value targets, while jammer drones interfere with defense interception systems. The collaborative work of the two types of drones increases the complexity of game decision-making. The existing technology does not take into account the collaborative work of the two types of drones, and therefore cannot make game decisions on the collaborative work. Summary of the invention
[0005] 1. Technical issues to be resolved
[0006] In view of the shortcomings of the prior art, the present invention provides an intelligent decision-making method and system for hybrid UAV formation collaborative game, which solves the technical problem that the prior art does not take into account the collaborative work of two types of UAVs.
[0007] (II) Technical solution
[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:
[0009] The present invention solves the technical problem by providing an intelligent decision-making method for a hybrid UAV formation collaborative game, the intelligent decision-making method being executed by a computer and comprising the following steps:
[0010] Acquire game participant data, wherein the game participant data includes attacker data and defender data; the attacker data includes attack drone information, jammer drone information, and attacker strategy set; the defender data includes defense target information, defense system information, and defender strategy set;
[0011] Calculating a strict game matrix based on the attacking strategy set and the defending strategy set, and obtaining an initial mixed strategy Nash equilibrium based on the strict game matrix;
[0012] Selecting an initial attacker strategy and an initial defender strategy according to the initial mixed strategy Nash equilibrium; the initial attacker strategy is the attacker strategy with the highest probability, and the initial defender strategy is the defender strategy with the highest probability;
[0013] Generate an attacker neighbor strategy set of the initial attacker strategy, and generate a defender neighbor strategy set of the initial defender strategy;
[0014] If the target attacker's neighbor strategy is not the initial attacker's strategy, and / or the target defender's neighbor strategy is not the initial defender's strategy, then the strict game matrix is updated to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix; the target attacker's neighbor strategy is the strategy that maximizes the attacker's profit value in the attacker's neighbor strategy set, and the target defender's neighbor strategy is the strategy that maximizes the defender's profit value in the defender's neighbor strategy set;
[0015] If the target attacker's neighbor strategy is the initial attacker strategy, and the target defender's neighbor strategy is the initial defender strategy, then the optimal attacker strategy is obtained according to the initial mixed strategy Nash equilibrium.
[0016] Preferably, it also includes:
[0017] Generate a zero-sum matrix game model, wherein the zero-sum matrix game model includes information of both parties in the game, a game strategy set, and a payoff matrix; the information of both parties in the game includes information of attacking drones, information of interfering drones, information of defense targets, and information of defense systems; the game strategy set includes an attacking party strategy set and a defending party strategy set;
[0018] The profit matrix includes:
[0019]
[0020] in,
[0021] For the attacking team in the strategic The following income value; is the tth strategy in the attacker’s strategy set, is the lth strategy in the defense strategy set; T is the attacking team, L is the defending team; S T is the set of attacking strategies, S L is the set of defense strategies;
[0022]
[0023] in,
[0024] represents the interception probability of defense system n against attacking drone m, P mk represents the probability of attack UAV m damaging the defense target k, v k represents the value of the defense target k, η n represents the defense coefficient of defense system n;
[0025] represents the probability that the jamming UAV g is not intercepted;
[0026] represents the effective interference coefficient of the jamming UAV g on the defense system n;
[0027] It represents the interception coefficient of the sub-defense system n after being interfered by the jamming UAV g;
[0028] It represents the interception coefficient after the sub-defense system n is actually interfered;
[0029] represents the probability that the attacking UAV m is not intercepted;
[0030] represents the effective damage probability of attack UAV m to defense target k;
[0031] represents the probability that the defense target k is not destroyed by the attacking UAV m;
[0032] represents the probability that the defense target k is not actually destroyed;
[0033] represents the probability that the defense target k is actually destroyed;
[0034] It represents the total expected damage value of the attacking formation to the defensive target.
[0035] Preferably, the calculating of the strict game matrix based on the attacking strategy set and the defending strategy set includes:
[0036] Selecting a plurality of attacker strategies from the attacker strategy set, and selecting a plurality of defender strategies from the defender strategy set;
[0037] Generate a strict game matrix according to the selected multiple attack strategies and multiple defense strategies;
[0038] The obtaining of the initial mixed strategy Nash equilibrium based on the strict game matrix includes:
[0039] The strict game matrix is processed based on a preset probability distribution algorithm to obtain an initial mixed strategy Nash equilibrium, wherein the initial mixed strategy Nash equilibrium includes a first probability distribution of multiple attacking strategies and a second probability distribution of multiple defending strategies.
[0040] Preferably, the selecting of the initial attacking strategy and the initial defending strategy according to the initial mixed strategy Nash equilibrium includes:
[0041] The initial mixed strategy Nash equilibrium is traversed, and the attacker strategy with the highest probability is obtained according to the first probability distribution, and is determined as the initial attacker strategy; the defender strategy with the highest probability is obtained according to the second probability distribution, and is determined as the initial defender strategy.
[0042] Preferably, generating the attacker neighbor strategy set of the initial attacker strategy includes:
[0043] Determine the game correspondence relationship in the initial attacker strategy, wherein the game correspondence relationship includes the attack correspondence relationship between the attacker's attack type UAV and the defender's defense target, and the interference correspondence relationship between the attacker's interference type UAV and the defender's defense system;
[0044] Select any one target game correspondence relationship in the initial attacking party strategy and determine the drone type of the drone in the target game correspondence relationship;
[0045] If the type of the drone is an attack drone, the defense target in the target game correspondence is changed to a neighbor defense target of the defense target, and a first attacker adjustment strategy is obtained; if the type of the drone is an interference drone, the defense system in the target game correspondence is changed to a neighbor defense system of the defense system, and a first attacker adjustment strategy is obtained;
[0046] Determining the first attacker adjustment strategy as an attacker neighbor strategy, and marking the initial attacker strategy as a first attacker adjustment strategy;
[0047] Selecting two game correspondences in the first attacking party adjustment strategy, and the drone types of the drones in the two game correspondences are the same;
[0048] Exchanging the defense targets or defense systems in the two game correspondences to obtain a second attacker adjustment strategy; determining the second attacker adjustment strategy as the attacker neighbor strategy;
[0049] An attacker neighbor strategy set is generated according to the initial attacker strategy and all attacker neighbor strategies.
[0050] Preferably, generating a defender neighbor strategy set for the initial defender strategy includes:
[0051] Determining a game correspondence relationship in the initial defender strategy, wherein the game correspondence relationship includes a defense correspondence relationship between the defense system and the attacker's drone;
[0052] Select any one of the game correspondence relationships in the initial defender strategy, and change the drone in the game correspondence relationship to a neighboring drone of the drone, to obtain a first defender adjustment strategy;
[0053] determining the first defender adjustment strategy as a defender neighbor strategy, and marking the initial defender strategy as a first defender adjustment strategy;
[0054] Select any two game correspondences in the first defender adjustment strategy, and exchange the drones in the two game correspondences to obtain a second defender adjustment strategy, and determine the second defender adjustment strategy as the defender neighbor strategy;
[0055] A defender neighbor strategy set is generated according to the initial defender strategy and all defender neighbor strategies.
[0056] Preferably, updating the strict game matrix comprises:
[0057] Based on the profit matrix, a target attacker neighbor strategy that maximizes the attacker's profit value is obtained from the attacker neighbor strategy set, and it is determined whether the target attacker neighbor strategy is the initial attacker strategy;
[0058] If not, then adding the target attacker's neighbor strategy to the strict game matrix, and re-performing the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix;
[0059] If so, then based on the payoff matrix, obtain the target defender neighbor strategy in the defender neighbor strategy set that maximizes the defender's payoff value, and determine whether the target defender neighbor strategy is the initial defender strategy; if not, add the target defender neighbor strategy to the strict game matrix, and re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix.
[0060] Preferably, the optimal attacking strategy is obtained according to the initial mixed strategy Nash equilibrium, including:
[0061] According to the first probability distribution, the attacker's strategy with the highest probability in the initial mixed strategy Nash equilibrium is obtained and determined as the best attacker's strategy.
[0062] The present invention solves the technical problem and provides an intelligent decision-making system for hybrid UAV formation collaborative game, comprising:
[0063] A data acquisition module is configured to acquire game participant data, wherein the game participant data includes attacker data and defender data; the attacker data includes attack drone information, jammer drone information, and attacker strategy set; the defender data includes defense target information, defense system information, and defender strategy set;
[0064] A mixed strategy Nash equilibrium acquisition module is configured to calculate a strict game matrix based on the attacking strategy set and the defending strategy set, and to acquire an initial mixed strategy Nash equilibrium based on the strict game matrix;
[0065] An initial strategy acquisition module is configured to select an initial attacker strategy and an initial defender strategy according to the initial mixed strategy Nash equilibrium; the initial attacker strategy is the attacker strategy with the highest probability, and the initial defender strategy is the defender strategy with the highest probability;
[0066] A neighbor strategy acquisition module is configured to generate an attacker neighbor strategy set of the initial attacker strategy and a defender neighbor strategy set of the initial defender strategy;
[0067] The optimal strategy module is configured to update the strict game matrix to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix if the target attacker neighbor strategy is not the initial attacker strategy and / or the target defender neighbor strategy is not the initial defender strategy; the target attacker neighbor strategy is the strategy that maximizes the attacker's profit value in the attacker neighbor strategy set, and the target defender neighbor strategy is the strategy that maximizes the defender's profit value in the defender neighbor strategy set;
[0068] If the target attacker's neighbor strategy is the initial attacker strategy, and the target defender's neighbor strategy is the initial defender strategy, then the optimal attacker strategy is obtained according to the initial mixed strategy Nash equilibrium.
[0069] (III) Beneficial effects
[0070] The present invention provides an intelligent decision-making method and system for hybrid UAV formation collaborative game. Compared with the prior art, it has the following beneficial effects:
[0071] The present invention obtains data of game participants, including attack drone information, jamming drone information, attacker strategy set, defense target information, defense system information, and defender strategy set. Calculate the strict game matrix, and obtain the initial mixed strategy Nash equilibrium based on the strict game matrix. Select the initial attacker strategy and initial defender strategy with the highest probability according to the initial mixed strategy Nash equilibrium. Generate the attacker neighbor strategy set of the initial attacker strategy, and generate the defender neighbor strategy set of the initial defender strategy. Calculate the mixed strategy Nash equilibrium with the maximum benefit through the neighbor strategy set, so as to obtain the best attacker strategy. Realize the game decision when the attack drone and the jamming drone work together. BRIEF DESCRIPTION OF THE DRAWINGS
[0072] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0073] Figure 1 A schematic diagram of a scenario of an intelligent decision-making method for hybrid UAV formation collaborative game provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0074] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0075] The embodiments of the present application solve the problem that the prior art is unable to make game decisions on the collaborative work of attack drones and jamming drones by providing an intelligent decision-making method and system for hybrid drone formation collaborative game.
[0076] The technical solution in the embodiment of the present application is to solve the above technical problems, and the overall idea is as follows:
[0077] The embodiment of the present invention obtains data of game participants, including attack drone information, jamming drone information, attacker strategy set, defense target information, defense system information, and defender strategy set. Calculate the strict game matrix, and obtain the initial mixed strategy Nash equilibrium based on the strict game matrix. Select the initial attacker strategy and initial defender strategy with the highest probability according to the initial mixed strategy Nash equilibrium. Generate an attacker neighbor strategy set of the initial attacker strategy, and generate a defender neighbor strategy set of the initial defender strategy. Calculate the mixed strategy Nash equilibrium with the maximum benefit through the neighbor strategy set, so as to obtain the best attacker strategy. Realize the game decision when the attack drone and the jamming drone work together.
[0078] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0079] An embodiment of the present invention provides an intelligent decision-making method for a hybrid UAV formation collaborative game, which is executed by a computer and includes the following steps:
[0080] S1. Obtain the data of the game participants, including the attacker data and the defender data; the attacker data includes the attack drone information, the jammer drone information and the attacker strategy set; the defender data includes the defense target information, the defense system information and the defender strategy set;
[0081] S2. Calculate a strict game matrix based on the attacking strategy set and the defending strategy set, and obtain an initial mixed strategy Nash equilibrium based on the strict game matrix;
[0082] S3. Select the initial attacker strategy and the initial defender strategy according to the initial mixed strategy Nash equilibrium; the initial attacker strategy is the attacker strategy with the highest probability, and the initial defender strategy is the defender strategy with the highest probability;
[0083] S4, generating an attacker neighbor strategy set of an initial attacker strategy, and generating a defender neighbor strategy set of an initial defender strategy;
[0084] S5. If the target attacker's neighbor strategy is not the initial attacker strategy, and / or the target defender's neighbor strategy is not the initial defender strategy, then update the strict game matrix to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix; the target attacker's neighbor strategy is the strategy that maximizes the attacker's payoff value in the attacker's neighbor strategy set, and the target defender's neighbor strategy is the strategy that maximizes the defender's payoff value in the defender's neighbor strategy set;
[0085] If the target attacker's neighbor strategy is the initial attacker strategy, and the target defender's neighbor strategy is the initial defender strategy, then the optimal attacker strategy is obtained according to the initial mixed strategy Nash equilibrium.
[0086] The following is a detailed analysis of each step.
[0087] In step S1, the data of the game participants are obtained, and a zero-sum matrix game model of multiple drones can be generated.
[0088] The zero-sum matrix game model includes information about both parties, a set of game strategies, and a payoff matrix. The information about both parties includes information about the attacking team and information about the defending team. The information about the attacking team includes information about attacking drones and jamming drones, and the information about the defending team includes information about defense targets and defense systems.
[0089] The game strategy set includes the attack strategy set of the attacking team and the defense strategy set of the defending team. Specifically, for the decision-making problem of multi-UAV coordinated ground attack, since the interception strategy of the defender will affect the benefits of the attack strategy of the attacker, it is modeled as a zero-sum matrix game model G = (P, S, U). The game model includes the participant part, the strategy part and the benefit part.
[0090] Among them, P = {T, L} is the participant of the game, and the two parties of the game include: T represents the attacker, and L represents the defender. The attacker includes multiple drones, which are divided into attack drones and interference drones. The defender includes multiple high-value targets and defense systems, and the defense system can launch multiple defense systems. High-value targets refer to units that the defender needs to defend, which can be units such as airports, roads, bridges, houses and radars. In the embodiment of this application, high-value targets are referred to as defense targets. It is assumed that the attacker has M attack drones and G interference drones. The defender has K high-value targets and N defense systems. Figure 1 A schematic diagram of a scenario of an intelligent decision-making method for hybrid UAV formation collaborative game provided in an embodiment of the present invention.
[0091] When the two sides are playing the game, the attacker uses attack drones to attack the defender's defense targets and uses jammer drones to interfere with the defender's defense system. In order to protect its own defense targets, the defender launches interceptor missiles through its defense system to intercept the incoming drones.
[0092] It should be noted that the defense system can intercept both attack drones and jammer drones, but it can only intercept one drone. An attack drone only attacks one defense target, and a jammer drone only interferes with one defense system.
[0093] Game strategies can be divided into the attack strategy of the attacker and the defense strategy of the defender. In the embodiment of the present application, the attack strategy is referred to as the attacker's strategy, and the defense strategy is referred to as the defender's strategy.
[0094] The attack strategy refers to the attack correspondence between the attack drones and the defenders during the game. For example, T has 2 attack drones and 1 jammer drone, and L has 5 defense targets and 3 defense systems. The attack strategy can be [1,5,3]. In the order of attack-jam, this attack strategy means that the 1st and 2nd attack drones of T attack the 1st and 5th defense targets of L respectively, and the 1st jammer drone of T interferes with the 3rd defense system of L.
[0095] The defender's strategy refers to the defensive correspondence between the defender's defense system intercepting the attacker's drones during the game. For example, when T has 4 drones and L has 3 defense systems, the defender's strategy [1,3,4] means that L's 1st, 2nd, and 3rd defense systems intercept T's 1st, 3rd, and 4th drones respectively.
[0096] It should be noted that since the defense system can intercept both attack drones and jammer drones, the defense strategy can also limit the type of drones to be defended. For example, the four drones of party T include three attack drones and one jammer drone, which can be numbered in the order of attack type-jamming type, with numbers 1-3 being attack type and number 4 being jammer type. Therefore, the defense strategy [1,3,4] means that the first and second defense systems of party L intercept the first and third attack drones (numbers 1 and 3) of party T respectively, and the third defense system of party L intercepts the first jammer drone (number 4) of party T.
[0097] In some embodiments, for a zero-sum matrix game model, let S≡S T ×S L The strategy space of the game. represents the set of attacker’s strategies, where τ T For S T The total number of strategies in For S T The t-th strategy in corresponds to the t-th attack strategy of party T. represents the defender’s strategy set, where τ L For S L The total number of strategies in For S L The lth strategy in the Lth strategy corresponds to the lth interception strategy of the L party.
[0098] Party T has a total of M drones and G jamming drones. Each attack drone can attack any high-value target of Party L, and each jamming drone can interfere with any sub-defense system of Party L. Party L has a total of K defense targets and N defense systems, and each defense system can intercept any drone of Party T. Then the number of strategies of Party T is τ T K M ·N G . Number of L-party strategies τ L is (M+G) N , that is, the size of the game strategy space is S|=τ T ·τ L =K M ·N G (M+G) N .
[0099] The payoff matrix includes:
[0100]
[0101] in,
[0102] For the attacking team in the strategic The profit value under Strategy execution The total expected damage value of the strategy to the defense targets of party L.
[0103] is the tth strategy in the attacker’s strategy set, is the lth strategy in the defender’s strategy set; S T is the set of attacking strategies, S L is the defense strategy set.
[0104] Due to the increased offensive and defensive relationship between the jamming UAV and the defense system, the concept of the defense coefficient of the defense system is introduced in the embodiment of the present application, which represents the effectiveness of the defense system in actual interception.
[0105]
[0106] in, represents the interception probability of defense system n against drone m, P mk represents the probability of damage of drone m to defense target k, v k represents the value of the defense target k, η n represents the defense coefficient of defense system n;
[0107] represents the probability that the jamming UAV g is not intercepted; represents the effective interference coefficient of the jamming UAV g on the defense system n; It represents the interception coefficient of the sub-defense system n after being interfered by the jamming UAV g; It represents the interception coefficient after the sub-defense system n is actually interfered; represents the probability that the attacking UAV m is not intercepted; represents the effective damage probability of attack UAV m to defense target k; represents the probability that the defense target k is not destroyed by the attacking UAV m;
[0108] represents the probability that the defense target k is not actually destroyed; represents the probability that the defense target k is actually destroyed; It represents the total expected damage value of the attacking formation to the defensive target.
[0109] It should be noted that for the same strategy A fixed profit value can be obtained, which is the profit value of the attacking formation. For the defending formation, its profit value is the negative of the profit value of the attacking formation.
[0110] In some embodiments, the attacker decides on the target allocation scheme for drones, assigning each drone a high-value target to be attacked, so as to maximize the expected damage value of the high-value targets. The defender decides on the defense system allocation scheme for the defense system, determining the drones intercepted by each defense system, so as to minimize the expected damage value of the high-value targets.
[0111] In step S2, a strict game matrix is calculated based on the attacking strategy set and the defending strategy set, and an initial mixed strategy Nash equilibrium is obtained based on the strict game matrix. Specifically, the following steps are included:
[0112] S201: Select multiple attacker strategies from an attacker strategy set, and select multiple defender strategies from a defender strategy set.
[0113] Specifically, p strategies may be selected from the attacking strategy set, and q strategies may be selected from the defending strategy set; wherein p and q are positive integers, and p≥2, q≥2.
[0114] When selecting, you can choose the pure strategy set S of the T side T Randomly select 2 attack strategies and add them to the pure strategy set S of L T Randomly select 2 defender strategies.
[0115] S202: Generate a strict game matrix according to the selected multiple attack strategies and multiple defender strategies.
[0116] Among them, a strict game matrix is generated according to p strategies and q strategies, and the strict game matrix includes p×q strategy pairs. p attacker strategies and q defender strategies can form a game matrix with p columns and q rows, which is called a strict game matrix. Taking p=q=3 as an example, the strict game matrix is as follows:
[0117]
[0118] Among them, the three attack strategies are The three second treatments are Thus, a 3×3 strict game matrix is obtained, which has 9 strategy pairs. Each strategy pair corresponds to a profit value, which is the profit value of the attacking team. The negative of the profit value is the profit value of the defending team.
[0119] Step S203: Processing the strict game matrix based on a preset probability distribution algorithm to obtain an initial mixed strategy Nash equilibrium, where the initial mixed strategy Nash equilibrium includes a first probability distribution of multiple attacker strategies and a second probability distribution of multiple defender strategies.
[0120] The preset probability distribution algorithm can be a Lemke–Howson (LH) algorithm. The LH algorithm first converts the game matrix into a linear inequality, then expresses the linear programming inequality through a coordinate system, converts it into a corresponding polyhedron, and finally finds the Nash equilibrium by iterating the rule on the equilibrium path. Other probability distribution algorithms can also be used in the embodiments of the present application as long as the probability of each strategy can be obtained.
[0121] It should be noted that two probability distributions are obtained: the first probability distribution of p strategies and the second probability distribution of q strategies. Among them, each of the p strategies of the attacker will correspond to a probability, and the sum of the probabilities of the p strategies of the attacker is 1. Similarly, each of the q strategies of the defender will correspond to a probability, and the sum of the probabilities of the q strategies of the defender is 1.
[0122] In step S3, an initial attacker strategy and an initial defender strategy are selected according to the initial mixed strategy Nash equilibrium. The initial attacker strategy is the attacker strategy with the highest probability, and the initial defender strategy is the defender strategy with the highest probability. Specifically, the following steps are included:
[0123] The initial mixed strategy Nash equilibrium is traversed, and the attacking strategy with the highest probability is obtained according to the first probability distribution, and is determined as the initial attacking strategy; the defending strategy with the highest probability is obtained according to the second probability distribution, and is determined as the initial defending strategy.
[0124] In step S4, an attacker neighbor strategy set of an initial attacker strategy is generated, and a defender neighbor strategy set of an initial defender strategy is generated. Specifically, the following steps are included:
[0125] S401: Generate an attacker neighbor strategy set of an initial attacker strategy, where the attacker neighbor strategy set includes multiple attacker neighbor strategies.
[0126] The steps to obtain the attacking neighbor strategy set are as follows:
[0127] S4011. Determine the game correspondence relationship in the initial attacking party's strategy, where the game correspondence relationship includes the attack correspondence relationship between the attacking party's attack-type UAV and the defending party's defense target, and the interference correspondence relationship between the attacking party's interference-type UAV and the defending party's defense system.
[0128] The attack correspondence relationship refers to which attack drone attacks which defense target, and the interference correspondence relationship refers to which interference drone interferes with which defense system.
[0129] The attack strategy of T can be expressed in the form of a one-dimensional array It is expressed as:
[0130]
[0131] The first M elements represent the target allocation scheme for M attack drones, k i ∈{1,...,K} represents the specific attack target of the i-th attack drone, which can be any high-value target from 1 to K. The last G elements represent the target allocation scheme of N jamming drones, n j ∈{1,...,N} represents the specific jamming target of the jth jamming UAV, which can be any defense system from 1 to N.
[0132] For example, when T has 3 attack drones and 2 jammer drones, and L has 5 defense targets and 8 defense systems, the strategy [1,3,5,2,7] means that T’s 1st, 2nd, and 3rd attack drones attack L’s 1st, 3rd, and 5th defense targets respectively, and T’s 1st and 2nd jammer drones jam L’s 2nd and 7th defense systems respectively.
[0133] Therefore, there are five corresponding game relationships: T-side No. 1 attack drone attacks L-side No. 1 defense target, T-side No. 2 attack drone attacks L-side No. 3 defense target, T-side No. 3 attack drone attacks L-side No. 5 defense target; T-side No. 1 jammer drone jams L-side No. 2 defense system, T-side No. 2 jammer drone jams L-side No. 7 defense system. All the corresponding game relationships in the initial attack strategy can be determined.
[0134] S4012. Select any target game correspondence relationship in the initial attacking party strategy and determine the drone type of the drone in the target game correspondence relationship.
[0135] Specifically, for the initial attacker strategy, each game correspondence relationship can be adjusted in turn to obtain several new strategies, which are referred to as first attacker adjustment strategies in the embodiment of the present application. Compared with the initial attacker strategy, only one game correspondence relationship has changed in each first attacker adjustment strategy.
[0136] S4013: If the drone type is an attack drone, the defense target in the target game correspondence is changed to the neighbor defense target of the defense target, and the first attacker adjustment strategy is obtained. If the drone type is an interference drone, the defense system in the target game correspondence is changed to the neighbor defense system of the defense system, and the first attacker adjustment strategy is obtained.
[0137] Among them, neighbor defense targets and neighbor defense systems refer to targets that are adjacent to each other in order of number. For example, for a defense target numbered 3, its neighbor defense targets are numbered 2 or 4. For another example, if there are 8 defense systems in total, then for a defense system numbered 1, its neighbor defense systems are numbered 2 or 8.
[0138] In the present application embodiment, by A high-value target attacked by an attack drone i ∈{1,...,K} or n j ∈{1,...,N} is transformed into an adjacent high-value target k i -1 or k i +1, or The sub-defense system that was interfered by a certain jamming drone j ∈{1,...,N} is transformed into an adjacent high-value target n j -1 or n j +1 for generating Neighbor strategy.
[0139] Taking strategies [1, 3, 5, 2, 7] as an example, we can select any game correspondence relationship, such as the first one “T side No. 1 attack drone attacks L side No. 1 defense target”.
[0140] We can first determine that the drone type of the corresponding relationship is an attack drone, so change the defense target attacked by the drone (numbered 1) to the neighbor defense target (numbered 2 or 8), and obtain [2,3,5,2,7] or [8,3,5,2,7].
[0141] Each game correspondence can be processed to obtain all the first attacker adjustment strategies. The initial attacker strategy [1,3,5,2,7] obtains all the first attacker adjustment strategies and the initial attacker strategy as follows:
[0142]
[0143] S4014: Determine the first attacker adjustment strategy as the attacker neighbor strategy, and mark the initial attacker strategy as the first attacker adjustment strategy. It should be noted that the first attacker adjustment strategy can be further expanded to increase the number of strategies.
[0144] S4015. Select two game correspondences in the first attacking party adjustment strategy, and the drone types of the drones in the two game correspondences are the same.
[0145] For all the first attacker adjustment strategies and initial attacker strategies obtained, the order of two elements in each strategy can be swapped, that is, the defense targets or defense systems in the two game correspondences can be swapped to obtain multiple second attacker adjustment strategies.
[0146] It should be noted that since attack drones can only attack defense targets and jammer drones can only jam defense systems, it is necessary to swap elements of the same type (defense targets or defense systems) to ensure the rationality of the strategy. Therefore, the two selected game correspondences need to ensure that the drone types are the same. For example, the game correspondence is selected as follows: T-side No. 1 attack drone attacks L-side No. 1 defense target, and T-side No. 2 attack drone attacks L-side No. 3 defense target.
[0147] S4016. Exchange the defense targets or defense systems in the two game correspondences to obtain a second attacker adjustment strategy. Determine the second attacker adjustment strategy as the attacker neighbor strategy. For example, in the strategy [1,3,5,2,7], 1 and 3 can be exchanged to obtain a second attacker adjustment strategy [3,1,5,2,7].
[0148] S4017: Generate an attacker neighbor strategy set according to the initial attacker strategy and all attacker neighbor strategies.
[0149] The initial attacker strategy [1,3,5,2,7] results in the attacker neighbor strategy set:
[0150]
[0151] S402: Generate a defender neighbor strategy set of the initial defender strategy, including:
[0152] S4021. Determine the game correspondence relationship in the initial defender's strategy, where the game correspondence relationship includes the defense correspondence relationship of the defense system to the attacker's drone.
[0153] The defense correspondence refers to which defense system intercepts which drone.
[0154] Represents the current strategy of party L in the form of a one-dimensional array It is expressed as:
[0155]
[0156] where t i ∈{1,...,M,...,M+G} represents the specific interception target of the i-th sub-defense system, which can be any attacking drone or any jamming drone of the attacker, and its value range is from 1 to M+G. i ∈{1,...,M}, it means intercepting the tth i Attack drone, when t i ∈{M,...,M+G}, it means intercepting the tth i -M jammer drones.
[0157] It should be noted that all drones can be uniformly numbered in the order of attack-interference. For example, when the T side has 4 attack drones and 3 interference drones, the drone numbers can be set to 1-7, where 1-4 represent attack drones and 5-7 represent interference drones. Assuming that the L side has 3 sub-defense systems and the T side has 4 attack drones and 3 interference drones, the defense strategy [1,3,6] means that the 1st, 2nd, and 3rd sub-defense systems of the L side intercept the 1st and 3rd attack drones and the 2nd interference drone (drone 6) of the T side respectively.
[0158] Therefore, there are three corresponding relationships in the game: L's No. 1 defense system intercepts T's No. 1 attack UAV, L's No. 2 defense system intercepts T's No. 3 attack UAV, and L's No. 3 defense system intercepts T's No. 2 interference UAV (the total order is No. 6 UAV).
[0159] S4022. Select any one game correspondence relationship in the initial defender strategy, and change the drone in the game correspondence relationship to a neighboring drone of the drone to obtain a first defender adjustment strategy.
[0160] Specifically, for the initial defender strategy, each game correspondence relationship can be adjusted in turn to obtain several new strategies, which are referred to as first defender adjustment strategies in the embodiment of the present application. Compared with the initial defender strategy, only one game correspondence relationship has changed in each first defender adjustment strategy.
[0161] When changing a drone in the game correspondence relationship to a neighbor drone, the change can be made in the order of all drones or in the order of drones of the current type.
[0162] For example, the defender’s strategy is [1,3,6]. If we select one of the game correspondences and change the drones in the order of 1-7, then the first defender’s adjustment strategy after all changes is: neighbor([1,3,6]) = {[1,3,6],[2,3,6],[7,3,6],[1,4,6],[1,2,6],[1,3,5],[1,3,7]}
[0163] If the order of the current type of drone is changed, the drone type needs to be determined first. For example, when the selected game correspondence is "L side No. 1 defense system intercepts T side No. 1 attack drone", the drone type is an attack drone, and the drone is changed from 1 to 4. The neighbor drone of No. 1 attack drone is No. 2 or 4. The adjustment strategy of the first defense party after all changes is:
[0164] neighbor([1,3,6])={[1,3,6],[2,3,6],[4,3,6],[1,4,6],[1,2,6],[1,3,5],[1,3,7]}
[0165] S4023: Determine the first defender adjustment strategy as the defender neighbor strategy, and mark the initial defender strategy as the first defender adjustment strategy.
[0166] S4024. Select any two game correspondences in the first defender adjustment strategy, and exchange the drones in the two game correspondences to obtain the second defender adjustment strategy, and determine the second defender adjustment strategy as the defender neighbor strategy.
[0167] For all the first defender adjustment strategies and initial defender strategies obtained, the order of any two elements in each strategy can be swapped, that is, the drones in the two game correspondences can be swapped to obtain multiple second defender adjustment strategies.
[0168] S4025: Generate a defender neighbor strategy set based on the initial defender strategy and all defender neighbor strategies. For example, the first defender adjustment strategy is:
[0169] neighbor([1,3,6])={[1,3,6],[2,3,6],[4,3,6],[1,4,6],[1,2,6],[1,3,5],[1,3,7]}
[0170] Then all the second defender adjustment strategies are obtained as follows:
[0171]
[0172] In step S5, if the target attacker's neighbor strategy is not the initial attacker's strategy, and / or the target defender's neighbor strategy is not the initial defender's strategy, then the strict game matrix is updated to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix; the target attacker's neighbor strategy is the strategy that maximizes the attacker's profit value in the attacker's neighbor strategy set, and the target defender's neighbor strategy is the strategy that maximizes the defender's profit value in the defender's neighbor strategy set;
[0173] If the target attacker's neighbor strategy is the initial attacker strategy, and the target defender's neighbor strategy is the initial defender strategy, then the optimal attacker strategy is obtained according to the initial mixed strategy Nash equilibrium. Specifically, the following steps are included:
[0174] Based on the profit matrix, a target attacker neighbor strategy that maximizes the attacker's profit value is obtained from the attacker neighbor strategy set, and it is determined whether the target attacker neighbor strategy is the initial attacker strategy.
[0175] If not, the target attacker's neighbor strategy is added to the strict game matrix to update the strict game matrix. Jump to step S203 to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix, thereby starting the next round of iterative processing.
[0176] If so, that is, the target attacker's neighbor strategy is the initial attacker's strategy. This means that the attacker's strategy meets the conditions and is already the best. It is also necessary to judge the defender's strategy and determine the best defender's strategy.
[0177] Based on the profit matrix, a target defender neighbor strategy that maximizes the defender's profit value can be obtained from the defender neighbor strategy set, and it can be determined whether the target defender neighbor strategy is the initial defender strategy.
[0178] If not, the target defender's neighbor strategy is added to the strict game matrix to update the strict game matrix, and jump to step S203 to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix, thereby starting the next round of iterative processing.
[0179] If yes, it means that both the attacker's strategy and the defender's strategy meet the conditions and are the best. At this time, the target attacker's neighbor strategy is the initial attacker's strategy, and the target defender's neighbor strategy is the initial defender's strategy. Then the best attacker strategy can be obtained according to the initial mixed strategy Nash equilibrium.
[0180] According to the first probability distribution, the attacker strategy with the highest probability in the initial mixed strategy Nash equilibrium can be obtained and determined as the best attacker strategy. After the best attacker strategy is obtained, the attacker formation can be controlled to execute the best attacker strategy.
[0181] For the defender, the defender's strategy with the highest probability in the initial mixed strategy Nash equilibrium can be obtained according to the second probability distribution and determined as the best defender's strategy so that the defender can execute it.
[0182] The embodiment of the present invention further provides an intelligent decision-making system for hybrid UAV formation collaborative game, the system comprising:
[0183] The data acquisition module is configured to acquire the data of the game participants, the game participant data includes the attacker data and the defender data; the attacker data includes the attack drone information, the jammer drone information and the attacker strategy set; the defender data includes the defense target information, the defense system information and the defender strategy set;
[0184] A mixed strategy Nash equilibrium acquisition module is configured to calculate a strict game matrix based on an attacker strategy set and a defender strategy set, and to acquire an initial mixed strategy Nash equilibrium based on the strict game matrix;
[0185] The initial strategy acquisition module is configured to select an initial attacker strategy and an initial defender strategy according to an initial mixed strategy Nash equilibrium; the initial attacker strategy is the attacker strategy with the highest probability, and the initial defender strategy is the defender strategy with the highest probability;
[0186] A neighbor strategy acquisition module is configured to generate an attacker neighbor strategy set for an initial attacker strategy and a defender neighbor strategy set for an initial defender strategy;
[0187] The optimal strategy module is configured to update the strict game matrix to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix if the target attacker neighbor strategy is not the initial attacker strategy and / or the target defender neighbor strategy is not the initial defender strategy; the target attacker neighbor strategy is the strategy that maximizes the attacker's profit value in the attacker neighbor strategy set, and the target defender neighbor strategy is the strategy that maximizes the defender's profit value in the defender neighbor strategy set;
[0188] If the target attacker's neighbor strategy is the initial attacker strategy, and the target defender's neighbor strategy is the initial defender strategy, then the optimal attacker strategy is obtained according to the initial mixed strategy Nash equilibrium.
[0189] It is understandable that the above-mentioned intelligent decision-making system provided in the embodiment of the present invention corresponds to the above-mentioned intelligent decision-making method. The explanations, examples, beneficial effects and other parts of the relevant contents can refer to the corresponding contents in the intelligent decision-making method of hybrid UAV formation collaborative game, which will not be repeated here.
[0190] In summary, compared with the prior art, the present invention has the following beneficial effects:
[0191] The embodiment of the present invention obtains the data of the game participants, calculates the strict game matrix, and obtains the initial mixed strategy Nash equilibrium based on the strict game matrix. The initial attacker strategy and the initial defender strategy with the highest probability are selected according to the initial mixed strategy Nash equilibrium. The attacker neighbor strategy set of the initial attacker strategy is generated, and the defender neighbor strategy set of the initial defender strategy is generated. The mixed strategy Nash equilibrium with the maximum benefit is calculated through the neighbor strategy set, so as to obtain the best attacker strategy. The game decision-making when the attack drone and the jamming drone work together is realized.
[0192] It should be noted that, through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment. In the specification provided here, a large number of specific details are explained. However, it can be understood that the embodiments of the present invention can be practiced without these specific details. In some instances, known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this specification.
[0193] In this article, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0194] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. An intelligent decision-making method for hybrid UAV formation collaborative game, the intelligent decision-making method is executed by a computer, It is characterized in that The following steps are involved: Acquire game participant data, wherein the game participant data includes attacker data and defender data; the attacker data includes attack drone information, jammer drone information, and attacker strategy set; the defender data includes defense target information, defense system information, and defender strategy set; Calculating a strict game matrix based on the attacking strategy set and the defending strategy set, and obtaining an initial mixed strategy Nash equilibrium based on the strict game matrix; Selecting an initial attacker strategy and an initial defender strategy according to the initial mixed strategy Nash equilibrium; the initial attacker strategy is the attacker strategy with the highest probability, and the initial defender strategy is the defender strategy with the highest probability; Generate an attacker neighbor strategy set of the initial attacker strategy, and generate a defender neighbor strategy set of the initial defender strategy; If the target attacker's neighbor strategy is not the initial attacker's strategy, and / or the target defender's neighbor strategy is not the initial defender's strategy, then the strict game matrix is updated to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix; the target attacker's neighbor strategy is the strategy that maximizes the attacker's profit value in the attacker's neighbor strategy set, and the target defender's neighbor strategy is the strategy that maximizes the defender's profit value in the defender's neighbor strategy set; If the target attacker's neighbor strategy is the initial attacker strategy, and the target defender's neighbor strategy is the initial defender strategy, then the optimal attacker strategy is obtained according to the initial mixed strategy Nash equilibrium.
2. The intelligent decision-making method according to claim 1, It is characterized in that Also includes: Generate a zero-sum matrix game model, wherein the zero-sum matrix game model includes information of both parties in the game, a game strategy set, and a payoff matrix; the information of both parties in the game includes information of attacking drones, information of interfering drones, information of defense targets, and information of defense systems; the game strategy set includes an attacking party strategy set and a defending party strategy set; The profit matrix includes: in, For the attacking team in the strategic The following income value; is the tth strategy in the attacker’s strategy set, is the lth strategy in the defense strategy set; T is the attacking team, L is the defending team; S T is the set of attacking strategies, S L is the set of defense strategies; in, represents the interception probability of defense system n against attacking drone m, P mk represents the probability of attack UAV m damaging the defense target k, v k represents the value of the defense target k, η n represents the defense coefficient of defense system n; represents the probability that the jamming UAV g is not intercepted; represents the effective interference coefficient of the jamming UAV g on the defense system n; It represents the interception coefficient of the sub-defense system n after being interfered by the jamming UAV g; It represents the interception coefficient after the sub-defense system n is actually interfered; represents the probability that the attacking UAV m is not intercepted; represents the effective damage probability of attack UAV m to defense target k; represents the probability that the defense target k is not destroyed by the attacking UAV m; represents the probability that the defense target k is not actually destroyed; represents the probability that the defense target k is actually destroyed; It represents the total expected damage value of the attacking formation to the defensive target.
3. The intelligent decision-making method according to claim 1, It is characterized in that The calculating of a strict game matrix based on the attacking strategy set and the defending strategy set includes: Selecting a plurality of attacker strategies from the attacker strategy set, and selecting a plurality of defender strategies from the defender strategy set; Generate a strict game matrix according to the selected multiple attack strategies and multiple defense strategies; The obtaining of the initial mixed strategy Nash equilibrium based on the strict game matrix includes: The strict game matrix is processed based on a preset probability distribution algorithm to obtain an initial mixed strategy Nash equilibrium, wherein the initial mixed strategy Nash equilibrium includes a first probability distribution of multiple attacking strategies and a second probability distribution of multiple defending strategies.
4. The intelligent decision-making method according to claim 3, It is characterized in that The selecting an initial attacker strategy and an initial defender strategy according to the initial mixed strategy Nash equilibrium includes: The initial mixed strategy Nash equilibrium is traversed, and the attacker strategy with the highest probability is obtained according to the first probability distribution, and is determined as the initial attacker strategy; the defender strategy with the highest probability is obtained according to the second probability distribution, and is determined as the initial defender strategy.
5. The intelligent decision-making method according to claim 1, It is characterized in that The generating of the attacker neighbor strategy set for the initial attacker strategy includes: Determine the game correspondence relationship in the initial attacker strategy, wherein the game correspondence relationship includes the attack correspondence relationship between the attacker's attack type UAV and the defender's defense target, and the interference correspondence relationship between the attacker's interference type UAV and the defender's defense system; Select any one target game correspondence relationship in the initial attacking party strategy and determine the drone type of the drone in the target game correspondence relationship; If the type of the drone is an attack drone, the defense target in the target game correspondence is changed to a neighbor defense target of the defense target, and a first attacker adjustment strategy is obtained; if the type of the drone is an interference drone, the defense system in the target game correspondence is changed to a neighbor defense system of the defense system, and a first attacker adjustment strategy is obtained; Determining the first attacker adjustment strategy as an attacker neighbor strategy, and marking the initial attacker strategy as a first attacker adjustment strategy; Selecting two game correspondences in the first attacking party adjustment strategy, and the drone types of the drones in the two game correspondences are the same; Exchanging the defense targets or defense systems in the two game correspondences to obtain a second attacker adjustment strategy; determining the second attacker adjustment strategy as the attacker neighbor strategy; An attacker neighbor strategy set is generated according to the initial attacker strategy and all attacker neighbor strategies.
6. The intelligent decision-making method according to claim 1, It is characterized in that The generation of the defender neighbor strategy set for the initial defender strategy includes: Determining a game correspondence relationship in the initial defender strategy, wherein the game correspondence relationship includes a defense correspondence relationship between the defense system and the attacker's drone; Select any one of the game correspondence relationships in the initial defender strategy, and change the drone in the game correspondence relationship to a neighboring drone of the drone, to obtain a first defender adjustment strategy; determining the first defender adjustment strategy as a defender neighbor strategy, and marking the initial defender strategy as a first defender adjustment strategy; Select any two game correspondences in the first defender adjustment strategy, and exchange the drones in the two game correspondences to obtain a second defender adjustment strategy, and determine the second defender adjustment strategy as the defender neighbor strategy; A defender neighbor strategy set is generated according to the initial defender strategy and all defender neighbor strategies.
7. The intelligent decision-making method according to claim 3, It is characterized in that The updating of the strict game matrix comprises: Based on the profit matrix, a target attacker neighbor strategy that maximizes the attacker's profit value is obtained from the attacker neighbor strategy set, and it is determined whether the target attacker neighbor strategy is the initial attacker strategy; If not, then adding the target attacker's neighbor strategy to the strict game matrix, and re-performing the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix; If so, then based on the payoff matrix, obtain the target defender neighbor strategy in the defender neighbor strategy set that maximizes the defender's payoff value, and determine whether the target defender neighbor strategy is the initial defender strategy; if not, add the target defender neighbor strategy to the strict game matrix, and re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix.
8. The intelligent decision-making method according to claim 3, It is characterized in that The step of obtaining the best attacking strategy according to the initial mixed strategy Nash equilibrium includes: According to the first probability distribution, the attacker's strategy with the highest probability in the initial mixed strategy Nash equilibrium is obtained and determined as the best attacker's strategy.
9. An intelligent decision-making system for hybrid UAV formation collaborative game, It is characterized in that The system comprises: A data acquisition module is configured to acquire game participant data, wherein the game participant data includes attacker data and defender data; the attacker data includes attack drone information, jammer drone information, and attacker strategy set; the defender data includes defense target information, defense system information, and defender strategy set; A mixed strategy Nash equilibrium acquisition module is configured to calculate a strict game matrix based on the attacking strategy set and the defending strategy set, and to acquire an initial mixed strategy Nash equilibrium based on the strict game matrix; An initial strategy acquisition module is configured to select an initial attacker strategy and an initial defender strategy according to the initial mixed strategy Nash equilibrium; the initial attacker strategy is the attacker strategy with the highest probability, and the initial defender strategy is the defender strategy with the highest probability; A neighbor strategy acquisition module is configured to generate an attacker neighbor strategy set of the initial attacker strategy and a defender neighbor strategy set of the initial defender strategy; The optimal strategy module is configured to update the strict game matrix to re-execute the step of obtaining the initial mixed strategy Nash equilibrium based on the strict game matrix if the target attacker neighbor strategy is not the initial attacker strategy and / or the target defender neighbor strategy is not the initial defender strategy; the target attacker neighbor strategy is the strategy that maximizes the attacker's profit value in the attacker neighbor strategy set, and the target defender neighbor strategy is the strategy that maximizes the defender's profit value in the defender neighbor strategy set; If the target attacker's neighbor strategy is the initial attacker strategy, and the target defender's neighbor strategy is the initial defender strategy, then the optimal attacker strategy is obtained according to the initial mixed strategy Nash equilibrium.
10. The intelligent decision-making system according to claim 9, It is characterized in that The generating of the attacker neighbor strategy set for the initial attacker strategy includes: Determine the game correspondence relationship in the initial attacker strategy, wherein the game correspondence relationship includes the attack correspondence relationship between the attacker's attack type UAV and the defender's defense target, and the interference correspondence relationship between the attacker's interference type UAV and the defender's defense system; Select any one target game correspondence relationship in the initial attacking party strategy and determine the drone type of the drone in the target game correspondence relationship; If the type of the drone is an attack drone, the defense target in the target game correspondence is changed to a neighbor defense target of the defense target, and a first attacker adjustment strategy is obtained; if the type of the drone is an interference drone, the defense system in the target game correspondence is changed to a neighbor defense system of the defense system, and a first attacker adjustment strategy is obtained; Determining the first attacker adjustment strategy as an attacker neighbor strategy, and marking the initial attacker strategy as a first attacker adjustment strategy; Selecting two game correspondences in the first attacking party adjustment strategy, and the drone types of the drones in the two game correspondences are the same; Exchanging the defense targets or defense systems in the two game correspondences to obtain a second attacker adjustment strategy; determining the second attacker adjustment strategy as the attacker neighbor strategy; An attacker neighbor strategy set is generated according to the initial attacker strategy and all attacker neighbor strategies.
Citation Information
Patent Citations
Unmanned aerial vehicle cluster combat game decision method based on quantum krill population evolution mechanism
CN109460056A
Multi-unmanned aerial vehicle cooperative game decision-making method and system
CN115328189A
Aircraft cluster confrontation method based on dynamic game and related device
CN116736889A
Optimal Strategies in Security Games
US20130273514A1
Cited By
Reinforcement learning method and system for intelligent decision making of unmanned aerial vehicle, and storage medium
CN119647631A
