Solving the Rush Hour Puzzle Problem by Different Heuristics
摘要
This research focuses on modeling and optimization of the Rush Hour puzzle, a grid-based board game whose objective is to determine a shortest sequence of movements of cars to let the red car exit a crowded parking lot. Recognized for its PSPACE-complete complexity, the Rush Hour problem presents significant challenges. The study explores Breadth-First Search (BFS) and A* search algorithms with various heuristics within the Subgoal framework. The implementations are evaluated against the 10,000 most complex Rush Hour configurations. Results demonstrate that the A* search algorithm markedly decreases the count of nodes explored during the solution-finding process. The heuristics developed in this study, along with the rush hour solver, provides possibilities for refining the subgoal search algorithm in a later step.