Flexible decision support tool for dynamic single machine scheduling problems

Abstract

ABSTRACT This paper addresses dynamic single machine scheduling problems. It proposes a flexible decision support tool based on a non-preemptive version of the previously-established overlapping load adjustment approach. The proposed tool adopts a two-phase proactive-reactive tactic where the first phase only characterizes the set of feasible schedules. In the second reactive phase, however, actual short-term scheduling takes place through a user-induced dynamic-event module, which progressively builds the schedule as the horizon rolls forward while ensuring the feasibility of the obtained schedule at all times. The performance of the tool is assessed using 500 randomly generated test instances of various sizes. The obtained results illustrate the flexibility of the tool as measured by the high number of potential schedule alternatives. Furthermore, the results highlight the tool’s capability of yielding a large number of high-quality schedules with respect to three performance measures.