| Summary: | 碩士 === 國立高雄第一科技大學 === 營建工程所 === 92 === Water gates are important facilities for safeguarding from flood-damage by
controlling drainage. In past few years, flood-damage which make the people losing property and life are caused by intensive rainfall, rapid flows and unworkable Watergates.
Due to the concept of lean government in Taiwan, water gates maintenance was
commissioned to the private contractors from 1999. Until 2004, the maintenance plan are still made by mental model and experience. This research developed an adaptive management system by Genetic Algorithms (GA) which assists in build maintenance plan.
Watergates maintenance is similar to Vehicle Routine Problems (VRP). The most
different part lies in the water gate distribution pattern, the distribution of water gates is almost linear. The distance between two water gates is quite short, so the maintenance plan has the character of sequential ordering.
The performance of the developed GA in this research fail to meet the expected
objective defined in the initial stage of this research, so the system has to be improved continuously. This research explored different ways to improve the GA developed in the first stage of the research. Finally an algorithm mimics the manual search characteristics was developed and it performs better than both the GA and manual method. It is also concluded that GA may not be an appropriate method to solve water gates routing problem.
|
|---|
