| Summary: | 碩士 === 國立成功大學 === 資訊管理研究所 === 94 === With the coming of globalization and privatization, the oil market in Taiwan becomes more and more competitive today than before. The best solution to maximize productivity and to minimize operating costs is explored by each petroleum corporation for keeping competitiveness. Therefore, the utilization ratio of oil tanks has already been an important issue and a suitable maintenance schedule of oil tanks has become the source of enterprise's profit and competitiveness. However, the objective of maintenance scheduling is not only to maximize the oil supply level, but also to minimize the maintenance costs. The complex conditions make domain experts difficult to arrange maintenance tasks properly. So the information techniques are proposed to support maintenance scheduling problem in this research.
Maintenance problem is essentially a NP-Hard problem. Ant Colony Optimization (ACO) algorithm has already succeeded in applying to solve various kinds of NP-Hard problems and is proved better performance than other classical heuristic algorithms. However, there are some drawbacks with ACO algorithm. In this thesis, we try to enhance the ACO algorithm after literature reviewing and then apply it to solve the multiple-objective maintenance scheduling problem of oil tanks to find out a Pareto solution set. This non-dominanced solution set will make decision makers easy to arrange the maintenance scheduling in the large number of solutions.
|
|---|
