论文部分内容阅读
提出一种新的交互式 Multi-Agent 遗传算法.该算法使固定在网格上的相邻智能体之间进行交叉、变异、死亡与再生操作和最优智能体本身进行自学习,来提高智能体的能量,从而使得算法获得较强的全局收敛能力和局部搜索能力.用户在每代进化中,只需选择感兴趣的个体,而不用评价每个个体的适应值,使得用户的评价操作变得简单易行.函数优化和服装设计的仿真实验表明算法能以较快的进化速度收敛,并使用户总评价次数减少,从而有效缓解用户的疲劳.
A new interactive Multi-Agent genetic algorithm is proposed in this paper. This algorithm can make the intelligent agents, such as crossover, mutation, death and regeneration operations, and the optimal agent self-learning, So that the algorithm obtains strong global convergence ability and local search ability.Users in each generation of evolution, just select the individuals who are interested, without having to evaluate the fitness of each individual, making the user’s evaluation of the operation of change It is easy to do.The simulation of function optimization and fashion design shows that the algorithm can converge at a faster rate of evolution and reduce the total number of users’ evaluation so as to effectively relieve the user’s fatigue.