Monte Carlo tree search with optimal computing budget allocation
Yunchuan Li, Michael C. Fu, Jie Xu · 2019
We analyze a tree search problem with an underlying Markov decision process, in which the goal is to identify the best action at the root that achieves the highest cumulative reward. We present a new tree policy that optimally allocates a limited computing budget to maximize a lower bound on the probability of correctly selecting the best action at each node. Compared to the widely used Upper Confidence Bound (UCB) type of tree policies, the new tree policy presents a more balanced approach to manage the exploration and exploitation trade-off when the sampling budget is limited. Furthermore, UCB assumes that the support of reward distribution is known, whereas our algorithm relaxes this assumption, and can be applied to game trees with mild modifications. A numerical experiment is conducted to demonstrate the efficiency of our algorithm in selecting the best action at the root.