Scheduling of unrelated parallel machines with limited server availability on multiple production locations: a case study in knitted frabrics

Abstract

This paper studies a complex variation of the parallel machine scheduling (PMS) problem, as encountered at a Belgian producer of knitted fabrics. The aim is to assign N J jobs to N M unrelated parallel machines, minimising a weighted combination of job lateness and tardiness. Jobs are assigned specific release, and due dates and changeover times are sequence dependent. Current literature is extended by including geographically dispersed production locations, which influence job due dates and objective function coefficients. Furthermore, the changeover interference due to limited availability of technicians is also studied in this paper. The scheduling problem is solved using a hybrid meta-heuristic, which combines elements from simulated annealing and genetic algorithms. This hybrid meta-heuristic is capable of solving real-scale scheduling problems of up to 750 jobs, 75 machines and 10 production locations within reasonable computation time. This hybrid scheduling procedure is extended with heuristic dispatching rules capable of reducing the impact of changeover interference by 23% on average compared to the random scenario, for the case where a single technician is expected to serve up to 12 machines.