Temporally Adaptive A* Algorithm on Time-Dependent Transportation Network

Traditional solutions to shortest path problems on time-varying transportation networks only use traffic information at definite moment so as to ignore the fact that the travel time through a link is dependent on the time to enter it. In this paper, the travel speed instead of the travel time on each link of road networks was modelled as a time-interval dependent variable, and a FIFOsatisfied computational function of the link travel time was then deduced. At last, a temporally adaptive A* shortest path algorithm on this FIFO network was presented, where the time factor was introduced into the evaluation function, and the Euclidean distance divided by the maximum possible travel speed was used as heuristic evaluator. An experiment on the real road network shows that the proposed algorithm is capable of foreseeing and bypassing those forthcoming traffic congestions, only with a cost of about 10 percent more computational time than the traditional algorithm. Furthermore, frequent path reoptimization caused by the traditional algorithm gets avoided effectively. * Corresponding author. Email: [email protected].