CN-119440034-B - Cluster distributed dynamic target distribution method based on preference alliance game

CN119440034BCN 119440034 BCN119440034 BCN 119440034BCN-119440034-B

Abstract

The invention discloses a cluster distributed dynamic target distribution method based on preference alliance gaming, which comprises the steps of defining a decision variable matrix to represent a matching relation between a target and an aircraft, defining target conflict constraint and target matching constraint conditions of the clustered aircraft, establishing a cluster communication topology, constructing a hitting efficiency function based on probability that the aircraft hits the target successfully, further designing an aircraft profit function, establishing an aircraft optimization performance index, designing a preference alliance gaming model based on the profit function and the optimization performance index of the aircraft, establishing a preference alliance gaming model based on a preference alliance gaming model and a maximum information gain algorithm, and designing a distributed dynamic target distribution algorithm so that the clustered aircraft can achieve a Darnsh stable partition through local information interaction under the communication topology. The invention can solve the problem of dynamic target allocation of large-scale clusters and has the advantages of small communication load, strong adaptability and the like.

Inventors

YANG HAO
NI YUAN
JIANG BIN
CHENG YUEHUA
YU ZIQUAN
Hua Yueyang

Assignees

南京航空航天大学

Dates

Publication Date: 20260512
Application Date: 20240929

Claims (3)

1. The cluster distributed dynamic target allocation method based on the preference alliance game is characterized by comprising the following steps of: (1) Defining a decision variable matrix to represent a matching relation between a target and an aircraft, defining target conflict constraint and target matching constraint conditions of a clustered aircraft, and establishing a clustered communication topology; (2) Constructing a hit efficiency function based on the probability of successful hit of the aircraft on the target, further designing an aircraft profit function, and establishing an aircraft optimization performance index; (3) Designing a preference-based target selection rule based on a profit function and an optimization performance index of the aircraft, and establishing a preference alliance game model; (4) Based on a preference alliance game model and a maximum information gain algorithm, designing a distributed dynamic target allocation algorithm, so that the clustered aircrafts reach a darashi stable partition through local information interaction under a communication topology; the step (1) of defining a decision variable matrix to represent the matching relationship between the target and the aircraft is as follows: Given a given Individual targets The air vehicle is put on the frame of the aircraft, < Defining a decision variable matrix , wherein, Characterization of aircraft And objects Is used for the matching relation of the (a), If not, the matching fails, M is the aircraft set; in order to set the objects in the set, , Corresponding to the real target, the target Corresponding to standby; The cluster aircraft target conflict constraint and target matching constraint conditions in the step (1) are as follows: Considering that each aircraft can hit only one target, a target conflict constraint is established: ; considering the limited range of impact and maneuverability of an aircraft, according to aircraft Coordinate information of (2) Target(s) Coordinate information of (2) No-fly zone Center coordinates of (c) Radius of coverage Establishing a target matching constraint: ; ; ; ; ; ; Wherein, the Combines the remote striking capability and the maneuverability of the aircraft, Is a positive constant which is used to control the temperature, Is a parameter characterizing the remote striking capability of an aircraft, Is a parameter characterizing the ability of an aircraft to maneuver; is an aircraft To the target Is a straight line distance of (2); is an aircraft With the object The distance that the line of (c) overlaps the no-fly zone, Is a no-fly zone Center coordinates to aircraft With the object Is provided with a vertical distance of the connection line of (c), Is an aircraft To the target A set of no-fly zones to be avoided; the cluster communication topology in the step (1) is as follows: Defining aircraft according to target matching constraints The set of real targets that can be selected is Establishing an undirected graph with the aircraft set as a node To characterize the communication topology of the aircraft clusters, undirected graph Is the adjacent matrix of (a) Wherein, if Then Otherwise Definition of Is an aircraft Is a neighbor set of (a); the implementation process of the step (2) is as follows: Aircraft with a plurality of aircraft body And objects The closer the distance is, the shorter the overlapping area between the two connecting lines and the no-fly zone is, the stronger the striking efficiency of the aircraft is, thus establishing a single aircraft For the target Is the striking efficiency function of (a), i.e. aircraft For the target The probability function of successful striking is as follows: ; defining a currently selected target Is composed of aircraft Target, object The successful hit benefit is obtained by aggregation Inner aircraft apportionment, aircraft Selecting a target The profit function of (2) is: ; Building an aircraft The local optimization objectives of (a) are as follows: ; Wherein, the To value, i.e., the benefit obtained after the target was successfully hit, Is an aircraft Is used for optimizing the performance index of the system, Is an aircraft The object of the selection is to provide, Is an aircraft A neighbor-selected target set; the probability of successful hit of the aircraft on the target in the step (2) further comprises the probability of successful hit of the aircraft on the target by a plurality of aircraft Probability of successful striking: ; Wherein, the For the current selection target Is a set of aircraft of the general type, For a single aircraft For the target Is successful; The implementation process of the step (3) is as follows: Definition symbol Characterization of aircraft Preference relation for alliance, if for target The method comprises the following steps: ; Then Representing an aircraft Compared to joining a federation Preference is given to joining the federation ; Based on this preference, the target selection rules for building the aircraft are set up by giving a coalition partition For any aircraft If and only if there is a target And is also provided with So that The aircraft will match the current target Is adjusted to target I.e. from alliances Adjust to alliance ; Based on aircraft collections Target set And establishing a preference alliance game model based on the preference-based target selection rule.
2. The method of claim 1, wherein the distributed dynamic target allocation algorithm in step (4) includes target selection and information transmission and information reception and target update, and wherein: the target selection and information transmission are as follows: 1) Introducing variables Characterization of aircraft The selected target is set with its initial value Introducing collections To avoid aircraft Repeatedly selecting the same target, and setting the initial value as ; 2) Judgment set If the set is empty, jumping to 3) if the set is empty, if the set is not empty, collecting Any one of the targets is selected If it meets Then describe the aircraft Compared to joining a federation Preference is given to joining the federation Thus will Assignment to And will From a collection Removing, namely jumping to 2); 3) Computing aircraft By alliance Adjust to alliance Post-optimization performance index An increased amount of (2) Will be Sent to aircraft Is a neighbor to all neighbors of (a); the information receiving and target updating are as follows: judging whether the performance index change information of the aircraft in the neighbor set is satisfied or not according to the received performance index change information If so, then the current aircraft is described Changing the target selection to Then, the performance index is increased more than the neighbor, and the iteration is ended based on the thought of the maximum information gain algorithm Changing the target selection to If not, it is indicated that there is a neighbor whose performance index is increased by more than or equal to that of the aircraft Then the iteration ends, the aircraft Still select the target When (1) When the amount is increased Is 0.
3. The method for distributing clustered distributed dynamic targets based on preference coalition gaming as set forth in claim 1 wherein targets are assigned to The probability of successful striking is set as 。

Description

Cluster distributed dynamic target distribution method based on preference alliance game Technical Field The invention belongs to the field of distributed task allocation, and particularly relates to a cluster distributed dynamic target allocation method based on preference alliance gaming. Background Clustered aircraft are a class of systems made up of a large number of different types of aircraft, of simple construction. Through functional complementation and cooperation among aircrafts, the clustered aircrafts can adapt to complex and diverse task scenes such as emergency rescue, reconnaissance and fight, and the like, the efficiency and success rate of task execution are improved, and the clustered aircrafts have wide application prospects. The target allocation is one of core modules of a clustered aircraft decision layer, is used as an action guide for executing tasks, and aims to realize matching among the aircraft and the targets by analyzing the information of the aircraft, the targets and the environment, and provides important basis for subsequent track planning and individual control of the aircraft. However, due to the large scale of clustered aircraft, individual failure of the aircraft is frequent, and it is difficult to rely on the operator to directly assign and adjust the targets of each aircraft. Therefore, the research on the cluster distributed dynamic target allocation method has important significance for enlarging the cluster scale and improving the cluster reliability. The cluster distributed dynamic target allocation method needs to meet the following requirements that (1) information is locally known, namely, an aircraft individual can only obtain information of part of related aircrafts and cannot grasp the information of the whole cluster. (2) And the individual independent decision is that the aircraft individual needs to select an optimal target according to the obtained local information and update the matched target. (3) The convergence of the algorithm, i.e. the target assignment result of the clusters can tend to be unchanged by target selection and target updating of the aircraft individuals. (4) The fault tolerance of the algorithm, namely when part of the aircraft individuals fail, whether the cluster can autonomously realize optimization and adjustment of the target distribution result of the cluster through independent decisions of the aircraft individuals or not. The existing target distribution method is difficult to meet the requirements, so the invention faces the requirements, and designs a cluster distributed dynamic target distribution method based on the preference alliance game theory. Disclosure of Invention The invention aims to provide a cluster distributed dynamic target distribution method based on preference alliance game, which can solve the problem of large-scale cluster real-time distributed target distribution, reduce cluster communication load and improve cluster reliability. The invention discloses a cluster distributed dynamic target allocation method based on preference alliance game, which comprises the following steps: (1) Defining a decision variable matrix to represent a matching relation between a target and an aircraft, defining target conflict constraint and target matching constraint conditions of a clustered aircraft, and establishing a clustered communication topology; (2) Constructing a hit efficiency function based on the probability of successful hit of the aircraft on the target, further designing an aircraft profit function, and establishing an aircraft optimization performance index; (3) Designing a preference-based target selection rule based on a profit function and an optimization performance index of the aircraft, and establishing a preference alliance game model; (4) Based on the preference alliance game model and the maximum information gain algorithm, a distributed dynamic target allocation algorithm is designed, so that the clustered aircrafts reach the darashi stable partition through local information interaction under the communication topology. Further, the defining a decision variable matrix to characterize the matching relationship between the target and the aircraft in the step (1) is: Given p (objects and M aircraft, p > M), a decision variable matrix A= [ alpha ij]m×(p+1) is defined, wherein object 0 corresponds to standby, alpha ij epsilon {0,1} characterizes aircraft i epsilon M and object And a ij =1 indicates that the matching is successful, otherwise, the matching is failed. Further, the clustered aircraft target conflict constraint and target matching constraint conditions in the step (1) are: Considering that each aircraft can hit only one target, a target conflict constraint is established: considering that the striking range and the maneuverability of the aircraft are limited, according to the coordinate information of the aircraft i epsilon M Coordinate information of object j epsilon PCentral coordinat