中文摘要：

将制造企业动态联盟合作伙伴的选择和组合抽象为多目标优化问题，提出改进的蚁群算法——“小生境蚁群算法”及“小生境信息差”的创新概念并对其进行优化求解，在正反馈环节中引人时变参数并利用经验信息和启发信息进行解算，从而有效地防止遗传算法中的“早熟”和基本蚂蚁算法中可能发生的“停滞”状态，获得选择合作伙伴多目标组合优化问题的最优解．

英文摘要：

Selection and combination of Dynamic Enterprise Alliance partners is abstracted as an optimizing problem of multi-targets. This paper presents a new concept of Ant Colony Algorithm, the Microhabitat Ant Colony Optimization algorithm and the Microhabitat Information, to optimize the solution. Time parameters under the positive feedbacks and the experiential information as well as heuristic information are adopted to avoid the deadlock of running into the precocity of Genetic Algorithms and the possible stagnation in the basic ant algorithm, therefore, the problem of muhi-target optimizing in selecting partners is solved.