Deep Reinforcement Learning-Based Multi-Agent System with Advanced Actor–Critic Framework for Complex Environment
The development of artificial intelligence (AI) game agents that use deep reinforcement learning (DRL) algorithms to process visual information for decision-making has emerged as a key research focus in both academia and industry. However, previous game agents have struggled to execute multiple comm...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-02-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/13/5/754 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850228227573809152 |
|---|---|
| author | Zihao Cui Kailian Deng Hongtao Zhang Zhongyi Zha Sayed Jobaer |
| author_facet | Zihao Cui Kailian Deng Hongtao Zhang Zhongyi Zha Sayed Jobaer |
| author_sort | Zihao Cui |
| collection | DOAJ |
| description | The development of artificial intelligence (AI) game agents that use deep reinforcement learning (DRL) algorithms to process visual information for decision-making has emerged as a key research focus in both academia and industry. However, previous game agents have struggled to execute multiple commands simultaneously in a single decision, failing to accurately replicate the complex control patterns that characterize human gameplay. In this paper, we utilize the ViZDoom environment as the DRL research platform and transform the agent–environment interactions into a Partially Observable Markov Decision Process (POMDP). We introduce an advanced multi-agent deep reinforcement learning (DRL) framework, specifically a Multi-Agent Proximal Policy Optimization (MA-PPO), designed to optimize target acquisition while operating within defined ammunition and time constraints. In MA-PPO, each agent handles distinct parallel tasks with custom reward functions for performance evaluation. The agents make independent decisions while simultaneously executing multiple commands to mimic human-like gameplay behavior. Our evaluation compares MA-PPO against other DRL algorithms, showing a 30.67% performance improvement over the baseline algorithm. |
| format | Article |
| id | doaj-art-372b7597abab459f9b787b43511ab9e2 |
| institution | OA Journals |
| issn | 2227-7390 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj-art-372b7597abab459f9b787b43511ab9e22025-08-20T02:04:36ZengMDPI AGMathematics2227-73902025-02-0113575410.3390/math13050754Deep Reinforcement Learning-Based Multi-Agent System with Advanced Actor–Critic Framework for Complex EnvironmentZihao Cui0Kailian Deng1Hongtao Zhang2Zhongyi Zha3Sayed Jobaer4College of Information Science and Technology, Donghua University, Shanghai 201620, ChinaCollege of Information Science and Technology, Donghua University, Shanghai 201620, ChinaCollege of Information Science and Technology, Donghua University, Shanghai 201620, ChinaCollege of Information Science and Technology, Donghua University, Shanghai 201620, ChinaCollege of Information Science and Technology, Donghua University, Shanghai 201620, ChinaThe development of artificial intelligence (AI) game agents that use deep reinforcement learning (DRL) algorithms to process visual information for decision-making has emerged as a key research focus in both academia and industry. However, previous game agents have struggled to execute multiple commands simultaneously in a single decision, failing to accurately replicate the complex control patterns that characterize human gameplay. In this paper, we utilize the ViZDoom environment as the DRL research platform and transform the agent–environment interactions into a Partially Observable Markov Decision Process (POMDP). We introduce an advanced multi-agent deep reinforcement learning (DRL) framework, specifically a Multi-Agent Proximal Policy Optimization (MA-PPO), designed to optimize target acquisition while operating within defined ammunition and time constraints. In MA-PPO, each agent handles distinct parallel tasks with custom reward functions for performance evaluation. The agents make independent decisions while simultaneously executing multiple commands to mimic human-like gameplay behavior. Our evaluation compares MA-PPO against other DRL algorithms, showing a 30.67% performance improvement over the baseline algorithm.https://www.mdpi.com/2227-7390/13/5/754deep reinforcement learningconvolution neural networkpartially observable Markov decision processmulti-agent system |
| spellingShingle | Zihao Cui Kailian Deng Hongtao Zhang Zhongyi Zha Sayed Jobaer Deep Reinforcement Learning-Based Multi-Agent System with Advanced Actor–Critic Framework for Complex Environment Mathematics deep reinforcement learning convolution neural network partially observable Markov decision process multi-agent system |
| title | Deep Reinforcement Learning-Based Multi-Agent System with Advanced Actor–Critic Framework for Complex Environment |
| title_full | Deep Reinforcement Learning-Based Multi-Agent System with Advanced Actor–Critic Framework for Complex Environment |
| title_fullStr | Deep Reinforcement Learning-Based Multi-Agent System with Advanced Actor–Critic Framework for Complex Environment |
| title_full_unstemmed | Deep Reinforcement Learning-Based Multi-Agent System with Advanced Actor–Critic Framework for Complex Environment |
| title_short | Deep Reinforcement Learning-Based Multi-Agent System with Advanced Actor–Critic Framework for Complex Environment |
| title_sort | deep reinforcement learning based multi agent system with advanced actor critic framework for complex environment |
| topic | deep reinforcement learning convolution neural network partially observable Markov decision process multi-agent system |
| url | https://www.mdpi.com/2227-7390/13/5/754 |
| work_keys_str_mv | AT zihaocui deepreinforcementlearningbasedmultiagentsystemwithadvancedactorcriticframeworkforcomplexenvironment AT kailiandeng deepreinforcementlearningbasedmultiagentsystemwithadvancedactorcriticframeworkforcomplexenvironment AT hongtaozhang deepreinforcementlearningbasedmultiagentsystemwithadvancedactorcriticframeworkforcomplexenvironment AT zhongyizha deepreinforcementlearningbasedmultiagentsystemwithadvancedactorcriticframeworkforcomplexenvironment AT sayedjobaer deepreinforcementlearningbasedmultiagentsystemwithadvancedactorcriticframeworkforcomplexenvironment |