Journal of Bionic Engineering ›› 2025, Vol. 22 ›› Issue (4): 1683-1701.doi: 10.1007/s42235-025-00721-9
Achievement of Fish School Milling Motion Based on Distributed Multi-agent Reinforcement Learning
Jincun Liu1,2,3,4；Yinjie Ren1,2,3,4；Yang Liu1,2,3,4；Yan Meng1,2,3,4；Dong An1,2,3,4；Yaoguang Wei1,2,3,4 #br#
- 1 National Innovation Center for Digital Fishery, ChinaAgricultural University, Beijing 100083, China
2 Key Laboratory of Smart Farming Technologies for AquaticAnimals and Livestock, Ministry of Agriculture and RuralAffairs, China Agricultural University, Beijing100083, China 3 Beijing Engineering and Technology Research Centrefor Internet of Things in Agriculture, China AgriculturalUniversity, Beijing 100083, China
4 College of Information and Electrical Engineering, ChinaAgricultural University, Beijing 100083, China
Achievement of Fish School Milling Motion Based on Distributed Multi-agent Reinforcement Learning
Jincun Liu1,2,3,4；Yinjie Ren1,2,3,4；Yang Liu1,2,3,4；Yan Meng1,2,3,4；Dong An1,2,3,4；Yaoguang Wei1,2,3,4
- 1 National Innovation Center for Digital Fishery, ChinaAgricultural University, Beijing 100083, China
2 Key Laboratory of Smart Farming Technologies for AquaticAnimals and Livestock, Ministry of Agriculture and RuralAffairs, China Agricultural University, Beijing100083, China 3 Beijing Engineering and Technology Research Centrefor Internet of Things in Agriculture, China AgriculturalUniversity, Beijing 100083, China
4 College of Information and Electrical Engineering, ChinaAgricultural University, Beijing 100083, China
摘要： In recent years, significant research attention has been directed towards swarm intelligence. The Milling behavior of fish schools, a prime example of swarm intelligence, shows how simple rules followed by individual agents lead to complex collective behaviors. This paper studies Multi-Agent Reinforcement Learning to simulate fish schooling behavior, over-coming the challenges of tuning parameters in traditional models and addressing the limitations of single-agent methods in multi-agent environments. Based on this foundation, a novel Graph Convolutional Networks (GCN)-Critic MAD-DPG algorithm leveraging GCN is proposed to enhance cooperation among agents in a multi-agent system. Simulation experiments demonstrate that, compared to traditional single-agent algorithms, the proposed method not only exhibits significant advantages in terms of convergence speed and stability but also achieves tighter group formations and morenaturally aligned Milling behavior. Additionally, a fish school self-organizing behavior research platform based on an event-triggered mechanism has been developed, providing a robust tool for exploring dynamic behavioral changes under various conditions.