Modeling and scheduling for manufacturing grid workflows using timed Petri nets

A manufacturing grid workflow can be viewed as the composition of activities that execute on heterogeneous, distributed, and job-dependent resources in a virtual organization to accomplish a specific manufacturing goal. Scheduling emerges as a key issue in manufacturing grid workflows since it assigns manufacturing resources to perform the required activities such that the optimal service is obtained. However, scheduling problems prove to be NP-hard due to the inherent combination explosion vices adhered to discrete event systems, such as the grid manufacturing. To crack such a hard nut, we focus on the formulation of a heuristic search algorithm based on timed Petri nets. When the activities involved in the manufacturing grid workflow can be executed by more than one resource, this scheduling algorithm selects the optimal one so as to minimize the makespan. Furthermore, the proposed scheduling policy in this paper aims to be dynamic rather than static such that the whole algorithm can be applied in a dynamic manufacturing grid environment. Both the theoretical and experimental results validate the legitimacy of the proposed approach.