| Summary: | 碩士 === 國立臺灣科技大學 === 工業管理系 === 91 === The main purpose of job scheduling is to derive a sequence of jobs for operation under limited resources such that the sequence is optimal for some predetermined criterion. However, most research in this area does not take system deterioration into account. In general, a production system gradually deteriorates due to usage and proper maintenance actions are required to stop and restore the system back to an operating condition. As a result, the optimal job sequence may not be performed as scheduled. Therefore, it is necessary to incorporate the maintenance actions into the derivation of an optimal schedule. In this thesis, a probability model is employed to describe the deterioration of a single machine and both the optimal maintenance policy and the optimal schedule are derived such that the makespan of jobs is minimized. A searching procedure is developed for the case when jobs are resumable after the interruption incurred by a maintenance action. For the non-resumable case, the scheduling problem is transformed into a parallel-machine scheduling problem and both a branch and bound algorithm and a heuristic algorithm are developed for deriving an optimal schedule. Finally, some numerical examples are given to evaluate the performance of the optimal schedule.
|
|---|
