The predecessor and the accounting algorithm speed up shortest path calculations in traffic routing applications

As traffic routing applications usually are heavily burdened due to the many requests, a low execution time of the shortest path algorithms is of uttermost importance. In this article two goal-directed heuristics are presented, which reduce this execution time. By guiding the search toward the destination and neglecting the less interesting areas of the network a remarkable speedup can be realized, especially in large networks like traffic networks. The predecessor algorithm makes use of local information in order to guide the search towards the destination, while the accounting algorithm additionally uses the path's history to direct the search in the right direction. Both algorithms have been tested on a European road network. It is shown experimentally that these heuristics indeed realize a speedup and are more accurate than most existing goal-directed heuristics.