[BibTeX] [RIS]
Boltzmann-based Exploration for Robust Decentralized Multi-Agent Planning
Tipo de publicação: Inproceedings
Citação: Nguyen2026ICAPS
Publication status: Published
Booktitle: Proceedings of the 36th International Conference on Automated Planning and Scheduling (ICAPS)
Volume: 36
Número: 1
Ano: 2026
Mês: June
Páginas: 404--408
Publisher: AAAI Press
Location: Dublin, Ireland
ISSN: 2334-0835
URL: https://ojs.aaai.org/index.php...
DOI: 10.1609/icaps.v36i1.42854
Resumo: Decentralized Monte Carlo Tree Search (Dec-MCTS) is widely used for cooperative multi-agent planning but struggles in sparse or skewed reward environments. We introduce Coordinated Boltzmann MCTS (CB-MCTS), which replaces deterministic UCT with a stochastic Boltzmann policy and a decaying entropy bonus for sustained yet focused exploration. While Boltzmann exploration has been studied in single-agent MCTS, applying it in multi-agent systems poses unique challenges. CB-MCTS is the first to address this. We analyze CB-MCTS in the simple-regret setting and show in simulations that it outperforms Dec-MCTS in deceptive scenarios and remains competitive on standard benchmarks, providing a robust solution for multi-agent planning.
Palavras-chave: Boltzmann exploration, decentralized planning, Monte Carlo tree search, multi-agent planning
Autores Nguyen, Nhat
Nguyen, Duong
Rizzo, Gianluca
Nguyen, Hung
Adicionado por: []
Total mark: 0
Anexos
  • Nguyen2026_ICAPS_Boltzmann_Exp...
Notas
    Tópicos