Abstract:This paper investigates a dynamic confrontation problem where a swarm of agents with weak capabilities confronts multiple targets with strong capabilities. In such hostile situations, the agents typically strive to survive by cooperating to offend the targets and defend against potential attacks from the targets. To enhance the capability of such multi-agent systems, it is necessary to develop an efficient mechanism in which each agent autonomously makes decisions and cooperatively confronts targets. In this paper, a cooperation and decision-making framework is proposed for a multi-agent system confrontation, which is comprised of target allocation, tactical decision-making, and swarm motion control algorithms. First, to address the exponential growth problem of possible behavioral interactions with increasing numbers of agents, a hedonic coalition formation game algorithm is designed for the agents forming disjoint coalitions corresponding to different targets, i.e., the agents cooperating to confront the targets in the form of coalitions. Then, during the stage of attacking or defending the targets, each agent interacts with coalition members through information exchange. To implement effective cooperative behavior and explainable autonomous decision-making, a fuzzy cognitive map is designed to fuse situational information and obtain decision reference information for each agent's tactical decision-making. Moreover, to design effective attack/defense tactics and strategies for each agent, the tactical pursuit point method is utilized to develop a tactical pursuit point for each agent based on the decision reference information. Finally, a swarm motion control algorithm, including decision-oriented and swarm behavior rules, is designed to drive each agent towards the assigned target. Simulation results show the effectiveness of the designed framework and algorithms. In the confrontation, each agent adjusts its strategies according to different situations, occupies advantageous positions, and accomplishes cooperative attack/defense strategies to reduce casualties.

Research on the Attacking Behavior Model of Combat Agent

How Effective Are the Prevailing Attack-Defense Models for Cybersecurity Anyway?

Behavior modeling method and its application in combat simulation

Analysis and Research of Attack Model

Simulation Model of Network Instruction and Defense Using Multi-Agent

Warfare CAWSOM Model

A Model for Multi-Agent Heterogeneous Interaction Problems

Attacking c-MARL More Effectively: A Data Driven Approach

A Novel Antagonistic Weapon-Target Assignment Model Considering Uncertainty and Its Solution Using Decomposition Co-Evolution Algorithm.

Driving-forces Model on Individual Behavior in Scenarios Considering Moving Threat Agents

Operational Capability of Unit Evaluation Based on MAS

Research on Systematically Describing Attacks

Research on an attack knowledge based enhanced at-tack tree model

Modeling Of Intentional Attack In Risk Analysis Methods

A cooperation and decision-making framework in dynamic confrontation for multi-agent systems

Chasing behavior between a motile swarm and a motile food

Modeling of Attacking and Defending Strategies in Situations with Intentional Threats

Adversarial Decision-Making for Moving Target Defense: A Multi-Agent Markov Game and Reinforcement Learning Approach

Effectiveness Evaluation Model of Moving Target Defense Based on System Attack Surface

Research and Effectiveness Analysis of Intelligent Unmanned Equipment Group Collaboration in Combat Simulation System

Collective Dynamic Behavior of Heterogeneous Multiple Agent Systems