Automatický pathfinding v mobilní hře

Abstract

The bachelor's thesis deals with the selection of a suitable search algorithm for a 2D video game. It uses Snake as an example of a such game. Afterwards, it focuses on selecting several search algorithms and their comparison. Among the criteria of the algorithms we compare are the ability to find a path for the snake, determine if it is the shortest possible path, and compare the effectiveness of the algorithm in comprasion to a real player. From our evaluation, it appears that we can't use certain search algorithms for the game Snake, and that the use of some algorithms is more efficient than others. From our experiments we can conclude that, the A* algorithm is more efficient than the BFS or DFS algorithm. The best algorithm for us then is the "perturbed Hamiltonian cycle" which, unlike other algorithms that we disscused, can win the Snake game every time.