Online make-to-order (MTO) lead-time scheduling: a probabilistic approach

This paper considers a lead time (or due date) scheduling problem faced by a manufacturer that produces customised products according to specific customer orders. It uses a single server model to design an effective, yet simple heuristic procedure for both sequencing and lead time determination decisions. Typically, a series of jobs with individual arrival times and weights must be sequenced on a single machine. With each arrival, a lead time must be determined, and the job completed before the promised due date. An on-line heuristic may be developed in certain situations, based on limited information about the future, with the objective of minimising the average (or total) weighted due date for large size instances. Computational testing shows that this heuristic is more effective than the conventional scheduling rules not only for the average weighted due date, but also for the average weighted lead time, for both stable systems and heavily loaded ones.