Deteriorating job scheduling problem in a job-shop manufacturing system by multi-agent system

Abstract

The scheduling problem is typically investigated in the literature considering the job-processing times fixed and constant values. In real manufacturing systems, the job-processing times can change due to several factors such as machine and cutting tool deterioration and maintenance issues. This article studies the integration of the deteriorating processing time problem in a multi-agent architecture that supports the scheduling activities in a job-shop manufacturing system. The main problem addressed in this research concerns the policy to restore the workstations to their initial state. It is observed that the state of the workstations and the performance measures of the manufacturing system enables the decision of when to perform the restore activity. It is proposed a policy based on a variable threshold level to decide on the restoration of the workstations. A simulation environment is developed to test the proposed approach to investigate several performance measures. The simulation results highlight how the proposed policy allows to keep a high level of performance reducing the number of restore activities.