交互式蚁群遗传算法

Interactive Ant Colony Genetic Algorithm

针对交互式遗传算法因种群规模小、进化代数少和用户评价负担重而性能偏低的问题,将蚁群优化引入遗传操作过程,从而提出一种使用精英保留策略的交互式蚁群遗传算法.从人机交互和降低用户评价负担的角度设计了算法结构,即以ACO信息素矩阵作为构造解的依据,替代遗传算法中的选择算子,然后将蚂蚁构造的解进行交叉和变异操作之后交给用户评价,而用户只需确定当前代哪个解是迄今历史最优解即可,而不用评价每个个体的具体适应值,并以迄今历史最优解更新信息素.利用函数优化和汽车造型设计进行实验,结果表明该算法具有较高优化性能,并能有效降低用户疲劳.

For the low performance of interactive genetic algorithm due to the small population size, less evolution generation and heav- y evaluation burden, an interactive ant colony genetic algorithm （iACGA） using elitist strategy is proposed,and ant colony optimization （ ACO） is introduced in genetic manipulation. From the perspective Of human-computer interaction and reducing the burden of user evaluation,the algorithm structure of iACGA is designed. Firstly,the solutions are structured according to the ACO pheromone matrix, and selection operator in genetic algorithm is replaced by ACO pheromone. After the implementation of crossover and mutation operation, the system output of the solutions are displaying to human, iACGA User.only needs to point out one optimal solution in current generation of history without specific fitness of each individual, and pheromone is updated accordingly. The experiments of function optimization and car styling design show that the algorithm has higher performance, and can effectively reduce user fatigue.