|
Abstract:
|
We consider a traffic flow model where the information about the actual travel time for each alternative route is not available when each driver performs route selection. For such a traffic flow model, we examine two routing methods to minimize the average travel time over all vehicles running in the model. One method tries to minimize the average travel time globally. It is assumed in this method that a central manager determines the routes of all vehicles. Since the number of combinations of vehicles' routes exponentially increases as the number of vehicles increases, we need an efficient combinatorial optimization technique. In this paper, we employ a genetic algorithm to search for a near-optimal route combination for all vehicles. In the other method, each driver tries to minimize his/her own travel time locally with no central manager. It is assumed in this method that each driver selects the route with the shortest estimated travel time among alternative routes. Each driver uses a neural network for the travel time estimation. Through computational experiments, we clearly demonstrate the characteristic features of each method.
|
