On the Analysis of Complex Backup Strategies in Monte Carlo Tree Search (2016)
Khandelwal, Piyush, Liebman, Elad, Niekum, Scott, Stone, and Peter
Over the past decade, Monte Carlo Tree Search (MCTS) and specifically Upper Confidence Bound in Trees (UCT) have proven to be quite effective in large probabilistic planning domains. In this paper, we focus on how values are backpropagated in the MCTS tree, and apply complex return strategies from the Reinforcement Learning (RL) literature to MCTS, producing 4 new MCTS variants. We demonstrate that in some probabilistic planning benchmarks from the International Planning Competition (IPC), select- ing a MCTS variant with a backup strategy different from Monte Carlo averaging can lead to substantially better results. We also propose a hypothesis for why different backup strategies lead to different performance in particular environments, and manipulate a carefully structured grid-world domain to provide empirical evidence supporting our hypothesis.
In Proceedings of The 33rd International Conference on Machine Learning, pp. 1319--1328, New York City, NY, USA, June 2016.

Slides (PDF)
Piyush Khandelwal Ph.D. Alumni piyushk [at] cs utexas edu
Elad Liebman Ph.D. Student eladlieb [at] cs utexas edu
Peter Stone Faculty pstone [at] cs utexas edu