A Gibbs sampling based coordination scheduling method with fixed I/O constraint

Abstract

Distributed production technology is the basic technology of modern production. Distributed production needs to consider the fixed input and output constraint, which is different from traditional centralized production. The objective of distributed production scheduling with fixed I/O constraint is to find the optimal scheduling route in production network. In this paper, we proposed the GSS (Gibbs Sampling Scheduling) method to schedule the schedulable jobs, according to a scheduling metric. GSS guarantees each device has opportunities to be allocated, to reduce the load on critical devices. The scheduling metric utilizes PCM (processing cost model) and TCM (transportation cost model) as parameters. The PCM represents the impact of schedulable jobs on the unscheduled jobs. The TCM represents the importance of devices’ positions in the production network. We designed experiments to analyze the parameters of the GSS, and carried out the test on makespan and load balance. These experiments can verify the effectiveness and efficiency of the proposed GSS method.