Abstract
This paper proposes a mathematical model of an optimal periodic schedule revision policy for jobshop scheduling where we detect delays of tasks and perform a periodic schedule revision at iT/M (i= 1, 2, …, M) for a schedule period T. We first overview the jobshop scheduling and then investigate the property of schedule delays. The long-run average cost per unit time of the schedule revision policy is, secondly, formulated under the assumption that the number N (t) of delayed tasks occurring over (0, t] follows a non-homogeneous Poisson process with mean value function H (t) . It is shown that there always exists a finite frequency M* minimizing the long-run average cost. Through computational experiments, finally, we discuss the characteristics of the proposed policy and its applicability to jobshop scheduling problems.
