Study on the shortest path algorithm based on fluid neural network of in-vehicle traffic flow guidance system

Abstract

The shortest path algorithm is critical for dynamic traffic assignment (DTA) and for the realization of route guidance in ITS. In order to implement the guidance function quickly and accurately, this paper introduces the fluid neural network (FNN) and develops a new parallel method based on FNN and genetic algorithm (GA) for route guidance. A sub-searching process and parameter optimization are employed to improve the performance of FNN. It is indicated by simulation that this method can be used to find the shortest route quickly from the original node to destination node in traffic networks.