摘要： 传统蚁群优化算法难以量化定性系统的优化指标。为此，提出一种交互式最大最小蚂蚁算法。将路径中的信息素限制在最大最小区间内，利用全局历史最优解进行信息素更新和用户评价，选择当前代最感兴趣的解，无需给出每个解的具体优劣数量值，以提高算法性能和降低用户疲劳。仿真实验结果表明，该算法具有较好的搜索能力和较快的收敛速度。

关键词: 蚁群优化, 人机交互, 汽车造型, 用户疲劳, 信息素, 定性系统

Abstract: A novel Interactive Max-Min Ant System(IMMAS) is proposed to overcome the weakness that conventional Ant Colony Optimization (ACO) algorithm can not effectively solve the problem of qualitative system whose optimization indices are unable or difficult to be quantificated. In order to enhance searching performance of the algorithm and reduce user fatigue, pheromone on the paths is limited to a maximum-minimum intervals and updated by the globally best solution, and IMMAS user only need select a mostly interesting individual of current generation, and need not evaluate quantization of every solution. In the test of the application to the car styling design, the proposed algorithm achieves good search ability and high convergence speed.

Key words: Ant Colony Optimization(ACO), human-computer interaction, car styling, user fatigue, pheromone, qualitative system

中图分类号: 

• TP312