With the increasing market features of globalization, service and customization, the way manufacturers conduct manufacturing business is changing. Under this background, Cloud Manufacturing (CMfg) emerges as a new networked manufacturing model. However, CMfg is immature in many aspects, especially in exception handling of service composition execution. Due to the complexity of the enterprise manufacturing process, there are a large number of unpredictable abnormal events in the dynamic open cloud manufacturing environment (such as user demand change, machine failure, etc.), so in order to ensure the smooth implementation of the service combination, it is indispensable to establish an effective service exception handling mechanism in CMfg. Moreover, when an exception occurs, in order to ensure the smooth execution of the downstream services after the exception point, the exception handling must satisfy the strict time constraints. To realize the exception-handing of service-composition with the strict deadline or strict time constraints, this paper proposes a service-composition exception adaptive adjustment model, considering the influences of the logistics transferring time and cost. And the occupied time of the cloud services and the valid replacement time range of the exception service are considered as the constraints in this model. In addition, the processing quality, the cost, and the quality of service are set as the optimal objectives. On the above basis, a service-composition exception handling adaptive adjustment (SCEHAA) algorithm based on the improved ant colony optimization algorithm (ACO) is proposed and applied to address the above model. Finally, to validate the performance of SCEHAA, a case study and the comparison experiment between SCEHAA and other algorithms (Particle Swarm Optimization and Artificial Bee Colony) are performed. The results show that the SCEHAA algorithm can perform the adaptive adjustment of the service-composition with strict time limit effectively, through the adaptive service execution path reconfiguration and has fast convergence effects.

随着全球化，服务和定制的市场特征不断增长，制造商开展制造业务的方式正在发生变化。在这种背景下，云制造（CMfg）成为一种新的网络化制造模型。但是，CMfg在许多方面都不成熟，尤其是在服务组合执行的异常处理中。由于企业制造过程的复杂性，动态开放云制造环境中存在大量不可预测的异常事件（如用户需求变化，机器故障等），因此为了确保顺利实施服务组合，在CMfg中建立有效的服务异常处理机制是必不可少的。而且，当发生异常时 为了确保异常点之后下游服务的顺利执行，异常处理必须满足严格的时间限制。为了实现具有严格期限或严格时间限制的服务组合异常处理，提出了一种服务组合异常自适应调整模型，并考虑了物流转移时间和成本的影响。该模型将云服务的占用时间和异常服务的有效替换时间范围视为约束。此外，将处理质量，成本和服务质量设置为最佳目标。在上述基础上，提出了一种基于改进蚁群算法（ACO）的服务组合异常处理自适应调整算法（SCEHAA），并将其应用于上述模型。最后，为验证SCEHAA的性能，进行了案例研究和SCEHAA与其他算法（粒子群优化和人工蜂群）的比较实验。结果表明，通过自适应的服务执行路径重构，SCEHAA算法可以在严格的时限内有效地进行服务结构的自适应调整，收敛速度快。进行了案例研究和SCEHAA与其他算法（粒子群优化和人工蜂群）的比较实验。结果表明，通过自适应的服务执行路径重构，SCEHAA算法可以在严格的时限内有效地进行服务结构的自适应调整，收敛速度快。进行了案例研究和SCEHAA与其他算法（粒子群优化和人工蜂群）的比较实验。结果表明，通过自适应的服务执行路径重构，SCEHAA算法可以在严格的时限内有效地进行服务结构的自适应调整，收敛速度快。