An Improved Approach towards Multi-Agent Pursuit–Evasion Game Decision-Making Using Deep Reinforcement Learning

Abstract

A pursuit–evasion game is a classical maneuver confrontation problem in the multi-agent systems (MASs) domain. An online decision technique based on deep reinforcement learning (DRL) was developed in this paper to address the problem of environment sensing and decision-making in pursuit–evasion games. A control-oriented framework developed from the DRL-based multi-agent deep deterministic policy gradient (MADDPG) algorithm was built to implement multi-agent cooperative decision-making to overcome the limitation of the tedious state variables required for the traditionally complicated modeling process. To address the effects of errors between a model and a real scenario, this paper introduces adversarial disturbances. It also proposes a novel adversarial attack trick and adversarial learning MADDPG (A2-MADDPG) algorithm. By introducing an adversarial attack trick for the agents themselves, uncertainties of the real world are modeled, thereby optimizing robust training. During the training process, adversarial learning was incorporated into our algorithm to preprocess the actions of multiple agents, which enabled them to properly respond to uncertain dynamic changes in MASs. Experimental results verified that the proposed approach provides superior performance and effectiveness for pursuers and evaders, and both can learn the corresponding confrontational strategy during training.

Pursuit and Evasion Strategy of a Differential Game Based on Deep Reinforcement Learning

“… Maolin Wang et al (2019) implemented the DDPG model to an open pursuit–evasion environment to learn the control strategy. Several researchers ( Lowe et al, 2020 ; Singh et al, 2020 ; Wan et al, 2021 ) proposed an actor-critic multi-agent DDPG algorithm to preprocess actions of multiple agents in the virtual environment.…”

A Word-Granular Adversarial Attacks Framework for Causal Event Extraction

“…It currently performs well on many decision-based problems. For example, in games or other fields, it has many applications [ 21 , 22 , 23 ]. At the same time, we believe that the application of reinforcement learning to NLP still has great potential.…”

An Ecological and Neural Argument for Developing Pursuit-Based Cognitive Enrichment for Sea Lions in Human Care

“…A further intriguing possibility, and one very much in line with the value of cognition as a real-time adaptation mechanism, would be to provide a fully automated, AI-piloted vehicle, either with preset swim patterns or guided by machine learning in aimed at increasing evasion time. Such automated pursuit/avoidance agents have been explored virtually in video game development and in embodied agents by the military (e.g., [ 69 , 70 ]). A target that gets better at evading the better the sea lion gets at capturing it would certainly present a dynamic and species-relevant enrichment challenge.…”