2026· IEEE Control Systems Letters· Vol 10, pp. 2263-2268· 0 citations· 15 references
Abstract
This letter investigates inverse reinforcement learning (IRL) for leader-follower multi-agent systems (LFMASs) over signed graphs, a scenario where existing graphical game frameworks fail due to their assumption of non-negative edge weights. We consider a practical setting where a learner LFMAS imitates the optimal behavior of an expert LFMAS without knowing the expert’s cost functions. A graphical apprentice game is formulated on signed graphs, and a distributed data-driven IRL algorithm is proposed, enabling the learner to recover equivalent reward weights and replicate the expert’s policies. Theoretical analysis guarantees convergence to the expert’s reward structure and closed-loop stability of the learner system. Simulation results validate the effectiveness of the proposed method.
NashDreamer is proposed, a principled MBRL framework for two-player zero-sum IIGs that introduces a centralized Multi-Agent Recurrent State-Space Model (MARSSM) that decouples environment dynamics from the effect of players's strategies on their individual observations.
LLM-based Multi-Agent Systems (MAS) achieve strong performance on complex reasoning tasks by coordinating multiple agents, but at the cost of substantial token consumption. Recent work on automatic topology design, ARG-Designer, has reframed this problem as autoregressive graph generation. However, its training objecti...
Poomphob Suwannapichat, B. Changaival, Caesar Wu et al.· 0 citations
A graph-based safe multi-agent reinforcement learning (MARL) framework for cooperative navigation with time-varying topology is presented, integrating a attention-based actor and a Graph Attention Network (GAT) centralized critic, enabling scale-insensitive policy learning under time-varying communication topologies.
Sizhe Xiao, Li-Jing Dong, Rui-Ting Bai et al.· 0 citations
This article investigates heterogeneous high-order nonlinear multiagent systems (MASs) under an event-triggered mechanism (ETM) and proposes an approximately optimal consensus protocol derived by solving the Hamilton-Jacobi-Isaacs (HJI) equation with the proposed actor-critic-identifier (ACI) learning framework.
Xing-Yu Gao, Zhi-Hong Liu, Choonki Ahn et al.· IEEE Transactions on Cyberne...· 0 citations
This paper investigates the distributed generalized Nash equilibrium (GNE) seeking problem for purely monotone games with set and shared affine inequality constraints over communication-constrained networks. A continuous-time projected dynamical system is developed by combining a regularization technique with a penalty...
Sheng-Li Du, Shuo Li, Tian-Li Xu et al.· IEEE Transactions on Signal...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
