Towards optimal resources allocation in cloud manufacturing: new task decomposition strategy and service composition model

Abstract

Task decomposition and service optimal composition are the two core processes in cloud manufacturing. Considering that task decomposition is the necessary preprocess for service composition, this paper firstly proposes a new task decomposition optimization method based on task-service matching. Specifically, tasks and services are described in a unified way. Then, the manufacturing task is preliminarily decomposed by using the matching mechanism between the tasks and services. Finally, considering the matching degree between tasks and services, and the horizontal competition and the vertical cooperation among candidate service sets, the preliminarily decomposed subtasks are reorganized to obtain tasks with appropriate granularity. Furthermore, utilizing these subtasks with appropriate granularity and the corresponding service sets, this paper proposes a new service composition model based on E-CARGO, aiming at the problem that the existing service composition models are lack of consideration for competitiveness and cooperation within and between candidate service sets. Concretely, the mapping relationship between the service composition and the E-CARGO model is firstly studied, and the service composition model is built. Afterwards, CPLEX is used to solve the optimal composition model. The practical case shows that this task decomposition strategy can solve the problem of disconnection between task decomposition and task-service matching without relying on more expert systems. In addition, the new service composition model based on E-CARGO is more flexible, which can easily model more variable factors, and use CPLEX to solve the optimal allocation scheme quickly and stably.