Scheduling to Optimize Due Date Performance under Uncertainty of Processing Times
MetadataShow full item record
AbstractThis thesis contributes to the theoretical and practical aspects of scheduling research. It is dedicated to the analysis of scheduling a set of jobs on a single machine when the jobs have uncertain processing times and conformance to the due date is the performance objective. The findings reveal that scheduling based on the point estimates of the processing times, when times are actually uncertain, may not lead to an optimal job sequence. As well, the decisions as to when each job should start on the machine may not be optimal. These decisions are important when costs are associated with both early and tardy completion of jobs. A stochastic scheduling methodology, based on sampling using simulation and optimization using evolutionary search, has been introduced. Results and behaviour have been evaluated and compared with single-machine deterministic scheduling, based on optimization using point estimates. Furthermore, the methodology has also been extended to the two-machine flow shop problem. Results confirm performance improvement using stochastic scheduling.
Bibliography: p. 119-122
Schulich School of Engineering