Learning Cooperative Multi-Agent Policies With Partial Reward Decoupling

Freed, Benjamin; Kapoor, Aditya; Abraham, Ian; Schneider, Jeff; Choset, Howie

doi:10.1109/lra.2021.3135930

Cited by 4 publications

(1 citation statement)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Value factorization approaches include de Witt et al, who introduced COMIX, which employs a decentralizable joint actionvalue function with a per-agent factorization, and showed significant improvement over MADDPG and VDN [97]. Freed et al proposed to rely on learned attention mechanisms to identify and decouple subsets of robots, which reduced the gradient estimator variances in a variety of motion/path planning tasks, showing improvements over independent learning and COMA [98].…”

Section: Motion and Path Planningmentioning

confidence: 99%

Distributed Reinforcement Learning for Robot Teams: A Review

Wang¹,

Damani²,

Wang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Purpose of review: Recent advances in sensing, actuation, and computation have opened the door to multi-robot systems consisting of hundreds/thousands of robots, with promising applications to automated manufacturing, disaster relief, harvesting, last-mile delivery, port/airport operations, or search and rescue. The community has leveraged model-free multi-agent reinforcement learning (MARL) to devise efficient, scalable controllers for multi-robot systems (MRS). This review aims to provide an analysis of the state-of-the-art in distributed MARL for multi-robot cooperation. Recent findings: Decentralized MRS face fundamental challenges, such as non-stationarity and partial observability. Building upon the "centralized training, decentralized execution" paradigm, recent MARL approaches include independent learning, centralized critic, value decomposition, and communication learning approaches. Cooperative behaviors are demonstrated through AI benchmarks and fundamental real-world robotic capabilities such as multi-robot motion/path planning. Summary: This survey reports the challenges surrounding decentralized model-free MARL for multi-robot cooperation and existing classes of approaches. We present benchmarks and robotic applications along with a discussion on current open avenues for research.

show abstract

Section: Motion and Path Planningmentioning

confidence: 99%