A user-induced proactive-reactive scheduling tool for dynamic flow-shop environments

Abstract

This paper proposes a flexible scheduling tool for dynamic flow-shop production systems that is based on a non-preemptive version of the overlapping load adjustment approach. The developed tool adopts a novel two-phase proactive-reactive dynamic scheduling mechanism. The first phase characterises a set of possible scheduling alternatives without actually building any detailed schedule. This phase follows the earliest loading approach via exploiting the jobs’ slack times across all machines. During the second phase, the dynamic build-up of schedules takes place in a progressive manner through a user-induced module that utilises graphical displays capturing in real-time the impact of unforeseen events on the ongoing schedule. Extensive experiments involving450 randomly generated test instances of three different sizes are carried out to assess the performance of the scheduling tool. The flexibility of the tool is assessed by the number of possible scheduling alternatives offered to the end-user. The results unveiled a high number of scheduling possibilities ranging from 155 schedules for the small size instances to 96,202 schedules for the large size instances. The relatively low number of tardy jobs (0.79 jobs on average) clearly exemplifies the robustness and quality of the scheduling alternatives despite the high number of potentially available schedules.