it also suffers from the overestimation problem occurring in directed acyclic graph.
This paper points out that df-pn(r) has a serious issue of underestimating proof and disproof numbers, while. Df-pn(r) cures it by ignoring proof and disproof numbers that may lead to infinite loops. However, df-pn has a notorious problem of infinite loops when applied to domains with repetitions. In this work, I am developing random and different greedy strategies to analyze the effectiveness of different factors such as pieces’ size, corner availability, and first-player turn.ĭepth-first proof-number search (df-pn) is powerful AND/OR tree search to solve positions in games. Various algorithms were discovered and created to develop winning strategy and AI against human opponents. Just like other strategy board game such as chess, Blokus contains definite strategic patterns that can be solved with computer algorithms. The player that scores the highest wins the game. Each player has pieces with different shapes and sizes that can be placed onto the board, where each block within a piece counts as one point. The ultimate goal of the game is to place as many pieces onto the board as a player can while blocking off the opponent’s ability to place more pieces onto the board. However, the players can only place a piece that touches at least one corner of their own pieces on the board. Blokus (officially pronounced as “Block us”) is an abstract strategy board game with transparent Tetris-shaped, color pieces that players are trying to place onto the board.
0 kommentar(er)
