Abstract: Manufacturing cloud service composition is a new technology to improve the utilization and achieve the appreciation of cloud manufacturing resources, and support the rapid development of cloud manufacturing industries. As the cloud manufacturing becomes more sophisticated, a large number of manufacturing cloud services with the same founctionalities and different quality of service appear. How to build a composite service through these manufacturing cloud services that not only can meet user’s quality demand, but also has optimal quality is a challenging problem. To solve this problem, this paper introduces the Collaborative Learning（CL） and elite retation mechanism to the Max-Min Ant System（MMAS）, constructs a new optimal algorithm with learning ability, and then applies this algorithm to solve the problem of otpimal manufacturing cloud service composition. The Simulation results validate the effectiveness of this algorithm.

Key words: cloud manufacturing, service composition, Quality of Service（QoS）, Collaborative Learning（CL）, Max-Min Ant System（MMAS）

摘要： 制造云服务组合是一种提高云制造资源利用率，实现制造资源增值的新技术, 对云制造产业的快速发展具有重要的支撑作用。随着云制造技术的日益成熟，网络上出现了大量具有相同制造功能和不同服务质量的制造云服务，如何通过这些制造云服务构建出既能满足用户制造需求，又具有最优服务质量的组合服务是云制造领域面临的难题。针对这一问题，将协作学习、变异和精英保留机制引入最大最小蚁群算法，构造了具有学习和变异能力的最大最小蚁群算法，并使用该算法求解服务质量感知的制造云服务优化组合问题。仿真实验结果验证了算法的有效性。

关键词: 云制造, 服务组合, 服务质量, 协作学习, 最大最小蚁群算法