A Multi-agent Cooperative Learning System with Evolution of Social Roles

Hou Y, Sun M, Zeng Y, Ong Y-S, Jin Y, Ge H, Zhang Q (2023)
IEEE Transactions on Evolutionary Computation: 1-14.

Zeitschriftenaufsatz | E-Veröff. vor dem Druck | Englisch
 
Download
Es wurden keine Dateien hochgeladen. Nur Publikationsnachweis!
Autor*in
Hou, Yaqing; Sun, Mingyang; Zeng, Yifeng; Ong, Yew-Soon; Jin, YaochuUniBi ; Ge, Hongwei; Zhang, Qiang
Abstract / Bemerkung
Recent developments in reinforcement learning have been able to derive optimal policies for sophisticated and capable agents, and shown to achieve human-level performance on a number of challenging tasks. Unfortunately, when it comes to multi-agent systems (MASs), complexities such as non-stationarity and partial observability bring new challenges to the field. Building a flexible and efficient multi-agent reinforcement learning algorithm capable of handling complex tasks has to date remained an open challenge. This paper presents an Multi-Agent learning system with the evolution of Social Roles (eSRMA). The main interest is placed on solving the key issues in the definition and evolution of suitable roles, and optimizing the policies accompanied by social roles in MAS efficiently. Specifically, eSRMA incorporates and cultivates role division awareness of agents to improve the ability to deal with complex cooperative tasks. Each agent is assigned a role module, which can dynamically generate roles based on the individuals’ local observations. A novel multi-agent reinforcement learning algorithm (MARL) is designed as the principal driving force that governs the role-policy learning process by a role-attention credit assignment mechanism. Moreover, a role evolution process is developed to help agents dynamically choose appropriate roles in decision-making. Comprehensive experiments on the StarCraft II micromanagement benchmark have demonstrated that eSRMA exhibits superiority in achieving higher learning capability and efficiency for multiple agents compared to the state-of-the-art MARL methods.
Erscheinungsjahr
2023
Zeitschriftentitel
IEEE Transactions on Evolutionary Computation
Seite(n)
1-14
ISSN
1089-778X
eISSN
1941-0026
Page URI
https://pub.uni-bielefeld.de/record/2978325

Zitieren

Hou Y, Sun M, Zeng Y, et al. A Multi-agent Cooperative Learning System with Evolution of Social Roles. IEEE Transactions on Evolutionary Computation. 2023:1-14.
Hou, Y., Sun, M., Zeng, Y., Ong, Y. - S., Jin, Y., Ge, H., & Zhang, Q. (2023). A Multi-agent Cooperative Learning System with Evolution of Social Roles. IEEE Transactions on Evolutionary Computation, 1-14. https://doi.org/10.1109/TEVC.2023.3268076
Hou, Yaqing, Sun, Mingyang, Zeng, Yifeng, Ong, Yew-Soon, Jin, Yaochu, Ge, Hongwei, and Zhang, Qiang. 2023. “A Multi-agent Cooperative Learning System with Evolution of Social Roles”. IEEE Transactions on Evolutionary Computation, 1-14.
Hou, Y., Sun, M., Zeng, Y., Ong, Y. - S., Jin, Y., Ge, H., and Zhang, Q. (2023). A Multi-agent Cooperative Learning System with Evolution of Social Roles. IEEE Transactions on Evolutionary Computation, 1-14.
Hou, Y., et al., 2023. A Multi-agent Cooperative Learning System with Evolution of Social Roles. IEEE Transactions on Evolutionary Computation, , p 1-14.
Y. Hou, et al., “A Multi-agent Cooperative Learning System with Evolution of Social Roles”, IEEE Transactions on Evolutionary Computation, 2023, pp. 1-14.
Hou, Y., Sun, M., Zeng, Y., Ong, Y.-S., Jin, Y., Ge, H., Zhang, Q.: A Multi-agent Cooperative Learning System with Evolution of Social Roles. IEEE Transactions on Evolutionary Computation. 1-14 (2023).
Hou, Yaqing, Sun, Mingyang, Zeng, Yifeng, Ong, Yew-Soon, Jin, Yaochu, Ge, Hongwei, and Zhang, Qiang. “A Multi-agent Cooperative Learning System with Evolution of Social Roles”. IEEE Transactions on Evolutionary Computation (2023): 1-14.

Link(s) zu Volltext(en)
Access Level
Restricted Closed Access

Export

Markieren/ Markierung löschen
Markierte Publikationen

Open Data PUB

Suchen in

Google Scholar