Efficient Monte Carlo Tree Search with Boltzmann Exploration for Optimal Planning
The core message of this paper is to introduce two new Monte Carlo Tree Search (MCTS) algorithms, Boltzmann Tree Search (BTS) and Decaying Entropy Tree Search (DENTS), that utilize Boltzmann exploration policies to efficiently plan and converge to the optimal policy, addressing the limitations of prior MCTS methods like UCT and MENTS.