Hybrid production-system control-architecture for smart manufacturing

Abstract

Highly customized products with shorter life cycles characterize the market today: the smart manufacturing paradigm can answer these needs. In this latter production system context, the interaction between production resources (PRs) can be swiftly adapted to meet both the variety of customers’ needs and the optimization goals. In the scientific literature, several architectural configurations have been devised so far to this aim, namely: hierarchical, heterarchical or hybrid. Whether the hierarchical and heterarchical architectures provide respectively low reactivity and a reduced vision of the optimization opportunities at production system level, the hybrid architectures can mitigate the limit of both the previous architectures. However, no hybrid architecture can ensure all PRs are aware of how orienting their behavior to achieve the optimization goal of the manufacturing system with a minimal computational effort. In this paper, a new “hybrid architecture” is proposed to meet this goal. At each order entry, this architecture allows the PRs to be dynamically grouped. Each group has a supervisor, i.e. the optimizer, that has the responsibility: (1) to monitor the tasks on all the resources, (2) to compute the optimal manufacturing parameters and (3) to provide the optimization results to the resources of the group. A software prototype was developed to test the new architecture design in a simulated flow-shop and in a simplified job shop production.