基于改进遗传算法的坦克大战游戏设计

Design of Tank War Game Based on Improved Genetic Algorithm

传统的遗传算法在面对规模较大且数据复杂的数据时,效率普遍比较低下。在游戏的运行过程中游戏AI面对的情况普遍比较复杂。因此传统的遗传算法并不适用于游戏开发。根据游戏实际情况,提出基于行为树的改进遗传算法。通过引入行为树对AI的行为进行分层从而减少遗传算法中数据的复杂度,并运用此技术制作一款简易的坦克大战游戏验证其可靠性。

The traditional genetic algorithm is generally inefficient in the face of large-scale and complex data.In the running process of the game,the situation faced by game AI is generally more complex.Therefore,the traditional genetic algorithm is not suitable for game development.Ac⁃cording to the actual situation of the game,this paper proposes an improved genetic algorithm based on behavior tree.The behavior tree is introduced to stratify the behavior of AI to reduce the complexity of data in genetic algorithm.And use it to make a simple tank war game.