摘要
协作问题一直是多智能体系统研究的关键问题之一,该文给出了用遗传算法来实现多智能体协作的一种方法。该方法利用遗传算法来解决当多智能体系统无法得到环境信息或得到这些信息代价过高时,如何有效地产生它们的协同运动。利用该方法,对三个智能体协作把箱子搬到目标点,然后改变目标点,让智能体继续完成协作任务进行计算机仿真,结果表明遗传算法在动态环境下实现多智能体协作方面的可行性和有效性。
Cooperation is one of the most important problems in multi agent system. This paper proposed a multi agent cooperation method based on genetic algorithm, which can solve the problem when a system can not get environment information or cost too much for getting the information. In a simulation experiment, three agents move a box to the destination cooperatively and change the destination when a task is completed, then let the agents to complete a new task has. Its result shows the feasibility and efficiency when genetic algorithm is used to multi agent cooperation in dynamic environment.
出处
《计算机仿真》
CSCD
2006年第3期123-125,共3页
Computer Simulation
关键词
遗传算法
多智能体协作
运动规划
Genetic algorithm
Multi agent cooperation
Athletics planning
