Anders Jonsson

The paper proposes a scalable, distributed approach for constrained Multi-Agent Reinforcement Learning by using local consensus over dual variables to ensure global constraint satisfaction without centralized coordination.

The paper introduces the Terminal Representation (TR), a novel, lower-dimensional, and structurally distinct formulation for encoding reward-weighted trajectories in RL that bypasses the need for eigendecomposition while retaining the benefits of existing representations.

The paper introduces Coordination Graphs for Constrained Multi-Agent Reinforcement Learning (CG-CMARL), a scalable framework that decomposes complex joint action spaces into pairwise regions to handle coordination and constraints efficiently.