標題: | SOLVING JAPANESE PUZZLES WITH LOGICAL RULES AND DEPTH FIRST SEARCH ALGORITHM |
作者: | Jing, Min-Quan Yu, Chiung-Hsueh Lee, Hui-Lung Chen, Ling-Hwei 資訊工程學系 Department of Computer Science |
關鍵字: | Japanese puzzle;Nonogram;Depth first search;Branch and bound |
公開日期: | 2009 |
摘要: | Japanese puzzle is one of logical games popular in Japan and Netherlands. Solving Japanese puzzle is a NP-complete problem. There are some related papers proposed. Some use genetic algorithm (GA), but the solution may be wrong. Some use depth first search (DFS) algorithm, which is an exhaustive search, the execution speed is very slow. In this paper, we propose a puzzle solving algorithm to treat these problems. Based on the fact that most of Japanese puzzle are compact and contiguous, some logical rules are deduced to paint some cells. Then, the DFS algorithm with the "branch and bound" scheme, which is used to do early termination for those impossible paths, is used to solve those undetermined cells. Experimental results show that our algorithm can solve Japanese puzzles successfully, and the processing speed is significantly faster than that of DFS. |
URI: | http://hdl.handle.net/11536/14289 |
ISBN: | 978-1-4244-4705-3 |
期刊: | PROCEEDINGS OF 2009 INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND CYBERNETICS, VOLS 1-6 |
起始頁: | 2962 |
結束頁: | 2967 |
Appears in Collections: | Conferences Paper |