Exploring search space trees using an adapted version of Monte Carlo tree search for combinatorial optimization problems

In this article we propose a heuristic algorithm to explore search space trees associated with instances of combinatorial optimization problems. The is based on Monte Carlo tree search, popular in game playing that used and represents the state-of-the-art for number games. Several enhancements are proposed make more suitable context. These exploit structure problem aim efficiently by pruning subtrees, using simulation policy, reducing domains variables eliminating dominated value assignments beam width. was implemented its components specifically tailored two problems: quay crane scheduling non-crossing constraints 0-1 knapsack problem. For first our surpasses results several new best solutions found benchmark set instances. second typically produces near-optimal slightly worse than results, but it needs only small fraction time do so. indicate competitive entirely different