The Research on Efficient Route Planning for Avoiding Flooded Regions

• Main Authors: Li, Cheng-Han, 李承翰
• Other Authors: Chang, Ya-Hui
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/06788451035592522719

碩士 === 國立臺灣海洋大學 === 資訊工程學系 === 102 === The disaster brought by heavy rain has become more and more serious in Taiwan, and it has been an important research issue to provide warning messages before flood. In the past research, we have built a flood forecasting system. It can identify the landmark which might be inundated. In this thesis, we further investigate the problem of routing planning for landmark rescuing. That is, we wish to quickly determine a route which starts from a given departure point, such as a landmark, avoids flooded areas, and gets to the destination. We propose two methods. The first one is called the “Baseline” method. It uses the R-tree index to perform spatial join between flooded areas and the minimum bounding rectangles of roads to identify the flooded roads. The Dijkstra algorithm then operates on those remaining unflooded roads to find the shortest path avoiding flooded areas.The second one is called the “Cloud” method. We first submit the departure point and the destination point to the Google Maps routing planning service to get an initial shortest path. We then identify those flooded roads within the path, and find nearby alternative intersections. Then, the departure point, alternative intersections, and the destination point are operated by the GSP (Generalized Shortest Path) algorithm to find an improved shortest path, and finally, the end points of each road within this path will be submitted to Google Maps again to get the final route. We have implemented these two methods and performed a series of experiments to compare their efficiency, route lengths, and the ratio of successfully getting a route without passing through flooded areas. The results show that the Baseline method has higher success rates, shorter route paths. However, when the roadnetwork is larger, its efficiency will decrease a lot. In contrast, the Cloud method performs quite fast even on a large dataset.