Location Selection of Urban Disaster Prevention Shelter Facilities Based on Multi-objective Programming

• Main Authors: CHANG, WEI, 張瑋
• Other Authors: YEN, KUO-CHI
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/jgnzg4

碩士 === 國防大學 === 運籌管理學系 === 107 === Cities are densely populated areas where the development of economic activities is concentrated, and if a major disaster occurs, the human or economic loss can be very serious. Therefore, the location selection of prevention shelter facilities will affect the evacuation efficiency and the implementation of disaster relief. As a result, the disaster prevention and transportation planning should be included in the overall policy implementation. This study takes urban residents point of views to discuss not only the satisfaction of evacuation time after disaster but also considers the effect of road risk after disaster and as well considers the government's point of view of setting up prevention shelter facilities costs. This research objective is to establish a multi-objective urban prevention shelter facilities base location model, so that the number of evacuees do not overflow the prevention shelter facilities capacity limit. In this paper, the single objective problem is solved by gene algorithm, and then the multi-objective gene algorithm is used to design the solution method that fits the multi-objective model. Furthermore, aiming at the problem that the solution result of random weight gene algorithm is susceptible to weight, a robustness random weight gene algorithm is proposed, and explain the validity and feasibility of the model by numerical analysis. Finally, the robustness random weight gene algorithm and the fixed weight gene algorithm are compared and analyzed. It can prove that the proposed algorithm has better performance and can provide useful information for further urban construction and planning policy making.