著者

Tai-YanWu Ching-NungLin Shi-JimYen

雑誌

ゲームプログラミングワークショップ2013論文集

巻号頁・発行日

pp.15-18, 2013-11-01

---

Minesweeper is a Partially Observable Markov Decision Process(POMDP) problem.How to gain more information from current situation affects future decision making significantly.This paper proposes a heuristic solver which outperforms all present minesweeper solvers includingpure heuristic and hybrid single player MCTS with heuristic approaches. With our methodology,more hidden information can be extracted from the same situation than previous methods. Fur-thermore, this solver is practical with efficient performance.