Rescheduling of unrelated parallel machines with job-dependent setup times under forecasted machine breakdown

Abstract

We address a rescheduling problem of unrelated parallel machines with job-dependent setup times where a machine breakdown is known in advance. Typical rescheduling methods usually re-assign or re-sequence jobs from a given schedule after machines break down. Recently, machine breakdowns can be forecasted with high accuracy before their actual occurrences from IoT sensors or artificial intelligence methods. We therefore define a new rescheduling problem in which jobs are re-assigned before machine breakdowns occur, and propose a mathematical programming model with three objective measures, makespan, stability and penalty cost. We then develop a simulated annealing (SA) algorithm combined with a fuzzy logic controller for adjusting the parameters in SA. We demonstrate the performance of the proposed algorithm with extensive experiments.