基于蚁群算法的动态联盟伙伴选择研究

Studies on Selecting Partners of Enterprises Dynamic Alliance Based on ACO

提出“小生境蚁群算法”(MACO),在利用正反馈的同时,引入时变参数来利用经验信息和启发信息,并在局部寻优时结合了小生境信息差的思想,从而有效地防止遗传算法中出现的“早熟”问题和蚂蚁算法中发生的“停滞”状态。把制造企业动态联盟合作伙伴的选择抽象为多目标优化的问题,并建立了优化选择目标函数。运用MACO解算选择合作伙伴的多目标问题,获得最优解。

The Microhabitat Ant Colony Optimization （MACO） algorithm was built up. When positive feedback being used, MACO made use of experience information and heuristic information with time parameters, and combined Microhabitat Information Differece when local being optimized, without running into the precocity of Genetic Algorithms and the stagnation of the basic ant algorithm. The selecting parmers of the manufacturing enterprises dynamic alliance being the problem of multi-target optimizing, the target function of optimizing selecting was created. The problem of multi-target optimizing of the selecting partners was sloved with MACO. And the optimized answer was achieved.