| Summary: | 碩士 === 朝陽科技大學 === 工業工程與管理系碩士班 === 98 === Most scheduling problems are NP-hard combinational optimization problems. That is, they are extremely complex and difficult to solve. In this thesis, the preemptive open shop scheduling problem with a given job completion sequence is investigated.
Without loss of generality, we assume that the jobs must be completed in the sequence C1≦C2≦…≦Cn. We first examine the problems with the objective of minimizing total completion time (ΣCi) and total tardiness (ΣTi). Both of them can be formulated as linear programs. It is shown that in some case the unforced delay of job completion may be helpful. In other words, it is not always advantageous to complete jobs as early as possible. Then, we develop two heuristics, HEU_F and HEU_S, for the problem whose objective is minimizing number of tardy jobs (ΣUi).
To evaluate the performance of the proposed heuristics, the solutions obtained are compared with optimal solutions from LINGO software. For consistent problems, heuristic HEU_F has an average relative percentage deviation of 3.07%. Also, only 8.04% of the tested instances are not optimally solved. On the other hand, heuristic HEU_S has an average relative percentage deviation of 1.83%, and only 0.70% of the tested instances are not optimally solved. For non-consistent problems, heuristic HEU_S perfroms worse. It has an average relative percentage deviation of 24.65%, and 14.95% of the tested instances are not optimally solved.
|
|---|
