Using Neural Network for Heuristic Search

• Main Authors: Hung Che Chen, 陳弘哲
• Other Authors: J. D. Wei
• Format: Others
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/53225226426640344651

碩士 === 長庚大學 === 資訊工程學系 === 100 === In computer graphics, the shortest path is the path with the least weight that connects one node to another. Problem solving by search can also be classified as searching the shortest path. Oversized search space corresponding to a complicated game mode always causes the problem solving processes inefficient, especially when the process needs to open numerous neighboring nodes. In this thesis, we consider the characteristic of neural networks and use it to design a nonlinear heuristic function and thus to improve the search of A* algorithm. Using the information from the known solution paths, we can adjust the target values of the states that were involved in the near best solutions. In doing this, we build a more accurate heuristic function and thus improve the search efficiency. We conducted a series of experiments on 3D map search and puzzle games. Experimental results prove that adaptable heuristics can reduce the states being retrieved and thus can make the problem solving process more effective.