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

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 game...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Kaifang Wan, Dingwei Wu, Yiwei Zhai, Bo Li, Xiaoguang Gao, Zijian Hu
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	pursuit–evasion multi-agent deep reinforcement learning decision-making adversarial learning MADDPG Science Q Astrophysics QB460-466 Physics QC1-999
Acceso en línea:	https://doaj.org/article/db85460494d841e789b8fd670265ccb1
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:db85460494d841e789b8fd670265ccb1
record_format	dspace
spelling	oai:doaj.org-article:db85460494d841e789b8fd670265ccb12021-11-25T17:29:37ZAn Improved Approach towards Multi-Agent Pursuit–Evasion Game Decision-Making Using Deep Reinforcement Learning10.3390/e231114331099-4300https://doaj.org/article/db85460494d841e789b8fd670265ccb12021-10-01T00:00:00Zhttps://www.mdpi.com/1099-4300/23/11/1433https://doaj.org/toc/1099-4300A 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.Kaifang WanDingwei WuYiwei ZhaiBo LiXiaoguang GaoZijian HuMDPI AGarticlepursuit–evasionmulti-agentdeep reinforcement learningdecision-makingadversarial learningMADDPGScienceQAstrophysicsQB460-466PhysicsQC1-999ENEntropy, Vol 23, Iss 1433, p 1433 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	pursuit–evasion multi-agent deep reinforcement learning decision-making adversarial learning MADDPG Science Q Astrophysics QB460-466 Physics QC1-999
spellingShingle	pursuit–evasion multi-agent deep reinforcement learning decision-making adversarial learning MADDPG Science Q Astrophysics QB460-466 Physics QC1-999 Kaifang Wan Dingwei Wu Yiwei Zhai Bo Li Xiaoguang Gao Zijian Hu An Improved Approach towards Multi-Agent Pursuit–Evasion Game Decision-Making Using Deep Reinforcement Learning
description	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.
format	article
author	Kaifang Wan Dingwei Wu Yiwei Zhai Bo Li Xiaoguang Gao Zijian Hu
author_facet	Kaifang Wan Dingwei Wu Yiwei Zhai Bo Li Xiaoguang Gao Zijian Hu
author_sort	Kaifang Wan
title	An Improved Approach towards Multi-Agent Pursuit–Evasion Game Decision-Making Using Deep Reinforcement Learning
title_short	An Improved Approach towards Multi-Agent Pursuit–Evasion Game Decision-Making Using Deep Reinforcement Learning
title_full	An Improved Approach towards Multi-Agent Pursuit–Evasion Game Decision-Making Using Deep Reinforcement Learning
title_fullStr	An Improved Approach towards Multi-Agent Pursuit–Evasion Game Decision-Making Using Deep Reinforcement Learning
title_full_unstemmed	An Improved Approach towards Multi-Agent Pursuit–Evasion Game Decision-Making Using Deep Reinforcement Learning
title_sort	improved approach towards multi-agent pursuit–evasion game decision-making using deep reinforcement learning
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/db85460494d841e789b8fd670265ccb1
work_keys_str_mv	AT kaifangwan animprovedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT dingweiwu animprovedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT yiweizhai animprovedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT boli animprovedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT xiaoguanggao animprovedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT zijianhu animprovedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT kaifangwan improvedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT dingweiwu improvedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT yiweizhai improvedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT boli improvedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT xiaoguanggao improvedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning AT zijianhu improvedapproachtowardsmultiagentpursuitevasiongamedecisionmakingusingdeepreinforcementlearning
_version_	1718412316615114752

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

Ejemplares similares