Pathfinding Strategy for Multiple Non-Playing Characters in 2.5 D Game Worlds

This paper investigates and determines the optimal pathfinding strategy for non-playing characters (NPCs) in 2.5D game worlds. Three algorithms, Dijkstra's, Best-first Search (BFS) and the A* algorithm using Manhattan distance, Euclidean distance and Diagonal distance heuristics, are tested under different interaction schemes and test environments consisting of different levels of obstacles. The result shows that the A* algorithm is the optimal algorithm under the Manhattan distance Heuristic. Our tests did not reveal significant difference among the cooperative, non-cooperative or competitive interaction schemes.