A decremental approach with the A∗ algorithm for speeding-up the optimization process in dynamic shortest path problems

Abstract

Dynamic (time-dependent) network routing has become important due to the deployment of advanced traveler information systems in navigation systems. We study the problem of speeding-up the shortest path in continuous-time dynamic networks without a priori knowledge of the link travel times. We apply the A∗ algorithm using the decremental approach to reduce the network size and speed-up the optimization process in dynamic shortest path problems. We utilize the weighted metric multidimensional scaling technique to define the potential function in the A∗ algorithm by converting the link travel costs (times) into distances. Moreover, we evaluate the performance of the decremental approach with respect to the CPU times and optimal costs by applying the A∗ algorithm and Dijkstra’s algorithm to different networks.