ABSTRACT

During the process of cloud manufacturing, various uncertainties in the real world could have a significant impact on the smooth execution of task, and could render the planned composite manufacturing service (CMS) inefficient or even ineffective. Therefore, this paper proposes an optimal selection method to enhance the robustness of CMS during the planning stage. Firstly, the structure of robust CMS is proposed by arranging the preferred and alternative services for each subtask, and a robust service composition and optimal selection (rSCOS) model of cloud manufacturing is constructed by defining the expected Quality of Service. Then, the gABC-GWO (guiding artificial bee colony – grey wolf optimisation) algorithm is proposed to solve the rSCOS model efficiently, in which three improvement strategies for ABC algorithm are designed according to the characteristics of GWO. Finally, two experiments are implemented and the results show that QoS of the preferred scheme of robust CMS is approximately 1.29% lower than that of CMS on average, while its robustness is improved by 1.81% and 13.14% depending on the two robustness indexes. Compared with other commonly-used intelligence optimisation algorithms, gABC-GWO algorithm possesses better search performance without significantly increasing time consumption, which makes it more suitable for solving rSOCS problems.

摘要

在云制造过程中，现实世界中的各种不确定性可能会对任务的顺利执行产生重大影响，并使计划中的复合制造服务（CMS）效率低下甚至无效。因此，本文提出了一种在计划阶段提高CMS鲁棒性的最优选择方法。首先，通过为每个子任务安排首选和替代服务来提出鲁棒CMS的结构，并通过定义预期的服务质量来构建云制造的鲁棒服务组合和最优选择（rSCOS）模型。然后，提出了gABC-GWO（指导人工蜂群-灰太狼优化）算法来有效求解rSCOS模型，根据GWO的特点，设计了三种ABC算法的改进策略。最后，进行了两个实验，结果表明，健壮的CMS首选方案的QoS平均比CMS的QoS低约1.29％，而健壮性则根据这两个健壮性指标分别提高了1.81％和13.14％。与其他常用的智能优化算法相比，gABC-GWO算法具有更好的搜索性能，而不会显着增加时间消耗，这使其更适合解决rSOCS问题。14％取决于两个稳健性指标。与其他常用的智能优化算法相比，gABC-GWO算法具有更好的搜索性能，而不会显着增加时间消耗，这使其更适合解决rSOCS问题。14％取决于两个稳健性指标。与其他常用的智能优化算法相比，gABC-GWO算法具有更好的搜索性能，而不会显着增加时间消耗，这使其更适合解决rSOCS问题。