基于多Agent遗传算法求解迷宫游戏

游戏设计的实质是设计一些相互关联的游戏规则.通过分析迷宫游戏元素之间的关系,抽象制定出一套游戏规则,并在此基础上,提出了一种以寻路个体为单个智能体的多Agent遗传算法.通过C＋＋进行模拟仿真,实验结果验证了所提优化算法的合理性;与传统的遗传算法相比,此算法在避免早熟的同时有很快的全局收敛性.

Solution for the Maze Game Based on Multi-Agent Genetic Algorithm

The essence of game design is to design some related rules of the game.On the basis of the analysis of relationships among various elements in the maze game,a set of rules of the game is worked out abstractly in the paper.Then a multi-agent genetic algorithm,in which the individual of finding the maze's path is as an agent,is proposed.At last it is simulated by C++ and compared with the traditional genetic algorithm.The simulation results indicate that the proposed optimization model is reasonable and this ...