Research On Dynamic Route Guidance Based On Drivers’ Route Choice Behavior

With the economy growing steadily, the problems of the city development, such as theincreasingly terrible traffic jams, frequent traffic accidents and accompanying trafficpollutions, have become the main problems that trouble people’s lives and the social harmony.Confronted with such severe traffic problems, the Intelligent Transportation System (ITS)arises at the historic moment. It provides services to drivers with advanced managementtechniques. As the core of ITS, the Dynamic Route Guidance System has become a hot areaof research. This paper studied the comprehensive factors impacting drivers’ route choice andput forward a new route choice algorithm, and proposed a dynamic route guidance modelbased on drivers’ route choice behavior by analyzing the features of drivers’ route choicebehavior.To meet the needs of the traffic dynamic changing of the dynamic route guidance system,we proposed a dynamic route guidance algorithm based on improved Hopfield neural network(HNN) and Genetic Algorithm (GA). Aiming at the inherent weakness of the Hopfield NeuralNetwork (HNN) easily falling into local optimum and slowly converging, we applied GeneticAlgorithm (GA) to the HNN to guarantee the global searching capability of the neuralnetwork. Also, dynamic learning rate is introduced to accelerate the convergence rate of thenetwork.Many factors, including the drivers’ behavior, traveling environment and road conditionsshould be comprehensively considered when studying modern traffic system. To establish amore accurate dynamic route guidance model, the features of drivers’ route choice behaviormust be fully considered. Based on the analysis of current research into drivers’ behavior, thispaper further studied how behavior-cognition impacted the route choice during the whole trip,presented a drivers’ route choice behavior-cognition model, and described thedecision-making mechanism of the model during drivers’ trip.In the simulation experiments, both the proposed algorithm and the model are tested and verified. The simulation experiments on the algorithm are divided into two parts. One is in theoliver30problem, our algorithm is compared with traditional ones, and it turns out that thealgorithm in this paper can easily gain a better solution. The other is in the MapX5.0+VC6.0environment, by simulating the real traffic situation, our algorithm shows its advantage ofgetting a path that well meets the needs of drivers. Then, simulation experiments are carriedout on the co-simulation platform of Routing Platform and Network Simulation Platform totest the drivers’ route choice behavior cognition model. According to the experimental results,it’s clear that the behavior-cognition model presented in this paper can properly provide anoptimal path concerning the features of drivers’ route choice behavior.