Study of event-driven and periodic rescheduling on a single machine with unexpected disruptions

Abstract

This paper studies the rescheduling problem of a single machine facing unexpected disruptions in order to determine which parameters can help reducing the negative impacts of these disruptions on schedule performance. A Genetic Algorithm (GA) is used to generate the initial schedule and the updated ones according to a reactive strategy. The performance of event-driven rescheduling and periodic rescheduling policies are compared in terms of total tardiness and total cost of rescheduling. Other factors that may affect rescheduling such as disruption time, disruption duration and number of disruptions are investigated. The sensitivity of results to both due date tightness and cost factor variation is tested. The results showed that the timing of the occurrence of disruption as related to scheduling horizon has a major effect on determining the best rescheduling policy. Event-driven policy is superior to other policies for short infrequent disruptions. It was found that the periodic policy is more appropriate for long and frequent disruptions.