Scale-Adaptive Balancing of Exploration and Exploitation in Classical Planning

TL;DR

This paper introduces GreedyUCT-Normal, a novel MCTS/THTS algorithm that adapts UCB1 to handle reward scale variations in classical planning, improving planning efficiency.

Contribution

It proposes a new UCB1-Normal bandit approach within MCTS/THTS, addressing reward scale issues and enhancing planning performance over existing methods.

Findings

GreedyUCT-Normal finds more plans with fewer node expansions.

The new algorithm outperforms Greedy Best First Search and previous MCTS/THTS algorithms.

Handling reward variance improves planning efficiency.

Abstract

Balancing exploration and exploitation has been an important problem in both game tree search and automated planning. However, while the problem has been extensively analyzed within the Multi-Armed Bandit (MAB) literature, the planning community has had limited success when attempting to apply those results. We show that a more detailed theoretical understanding of MAB literature helps improve existing planning algorithms that are based on Monte Carlo Tree Search (MCTS) / Trial Based Heuristic Tree Search (THTS). In particular, THTS uses UCB1 MAB algorithms in an ad hoc manner, as UCB1's theoretical requirement of fixed bounded support reward distributions is not satisfied within heuristic search for classical planning. The core issue lies in UCB1's lack of adaptations to the different scales of the rewards. We propose GreedyUCT-Normal, a MCTS/THTS algorithm with UCB1-Normal bandit for agile classical planning, which handles distributions with different scales by taking the reward variance into consideration, and resulted in an improved algorithmic performance (more plans found with less node expansions) that outperforms Greedy Best First Search and existing MCTS/THTS-based algorithms (GreedyUCT,GreedyUCT*).