Tag
Mesh-RL is a spatial domain-decomposition framework for reinforcement learning that partitions the environment into overlapping subgrids to accelerate temporal-difference learning and long-range credit assignment, improving convergence speed and sample efficiency in sparse-reward environments.