An efficient shortest path computation system for real road networks

In this paper, we develop an efficient system to compute shortest paths in real road networks. An optimal shortest path algorithm is proposed based on a two-level hierarchical network structure. A pre-computation technique is used to improve the time efficiency. To further improve time efficiency and reduce memory requirement, we propose an algorithm to minimize the number of boundary nodes by relocating them between adjacent sub-networks. The performances of our approach with different network partition methods (with or without minimization of the number of boundary nodes) are compared in terms of both time efficiency and memory requirement. The experimental results on the real road network of Hong Kong demonstrated the efficiency of our method and the usefulness of minimizing of the number of boundary nodes.