Study on Near-Optimal Path Finding Strategies in a Road Network

Abstract

In this paper we analyze the existing defects of hierarchical and nonhierarchical path finding algorithm respectively in detail and present a hierarchical path finding algorithm based on the heuristic directing search approach which is used for promoting the nodes of lower level to higher level in the road network. The improved heuristic-hierarchical path finding algorithm was tested on road network of Xi'an city with a two-level hierarchy. The experimental results show that the heuristic based hierarchical path finding algorithm has been found to be over 25 times faster than the Dijkstra's algorithm and the ratio of the highest level road obtained by heuristic based hierarchical path finding algorithm is twice as much as the Dijkstra's algorithm. This means that the hierarchical path finding algorithm is suitable for the status of long-distance between origin and destination ( O- D). We also propose a novel yet simple path finding strategy based on a threshold, that is, if the distance of O- D is less than the threshold which was obtained by means of statistical analysis the nonhierarchical path finding algorithm will be used, otherwise we use hierarchical path finding algorithm to find a near-optimal or satisfactory path in city road network. The testing results of Xi'an city road network show that the roads which were found out using this new path finding strategy are easy to be adopted by drivers.