Summary: | 碩士 === 國立東華大學 === 資訊工程學系 === 101 === Minesweeper is a popular one player computer game. With simple rules, sometimes it is difficult to be solved. The main reason is that it is required to be solved by probability. However, there are few Chinese papers in this subject. The principal purposes of this study are to analysis of Minesweeper and to provide a more powerful and faster solving method with heuristics. In the beginning of this paper, introduce the rule and history of Minesweeper. Secondly, discuss strategy of human player in depth. Thirdly, design a heuristic and optimization method to solve Minesweeper.
In the previous method of solving Minesweeper, Constraint Satisfaction Problem method is a common solution which counts probability of mine for every grid exactly, so it has great win rate. In this paper, our method can count mine probability as well as number probability and use evaluation function to get evaluate value to find better choice. We show the better win rate in experiment.
|