Optimal Player Information MCTS applied to Chinese Dark Chess

Alpha-beta and Monte-Carlo Tree Search (MCTS)are two powerful paradigms useful in computer games. Whenconsidering imperfect information, the tree that represents thegame has to deal with chance. When facing such games thatpresents increasing branching factor, MCTS may consider, asalpha-beta do, pruning to keep efficiency. We present a modifiedversion of MCTS-Solver algorithm, called OPI-MCTS as OptimalPlayer Information MCTS, that adds game state informationto exploit logical reasoning during backpropagation and thatinfluences selection and expansion. OPI-MCTS is experimentedin Chinese Dark Chess, which is a imperfect information game.OPI-MCTS is compared with classical MCT