A Reinforcement Learning approach for the Cross-Domain Heuristic Search Challenge
Luca Di Gaspero, Tommaso Urli · 2011
The International Cross-Domain Heuristic Search Challenge (hereinafter CHeSC 2011) [3] is an ongoing competition that prompts for the design of a generally applicable high-level strategy for the automatic selection of problem-specific low-level heuristics across different problem domains. We participate in the challenge with a Reinforcement Learning approach. In this paper we describe the current state of the general algorithm by outlining our design choices and we present the preliminary results achieved by this approach. 1