Genetic and simulated annealing algorithms-based traffic state identification

Abstract

Accurate and scientific traffic state identification is the basis of traffic navigation system, traffic control system and traffic organization optimization. In this paper, the dynamic traffic data collected by geomagnetic detector are firstly used to identify traffic state. We proposed SAGA-FCM clustering algorithm which is combined simulated annealing algorithm (SA) with genetic algorithm (GA) for urban expressway traffic state identification. This method can overcome the problems those the dynamic data of other detectors are not accurate and the time interval is not uniform. Meanwhile, it can overcome the instability of FCM algorithm clustering center and it is easy to fall into the local extreme value and "premature" problem of genetic algorithm. Research results show that, Compared with FCM algorithm and GA-FCM algorithm, SAGA-FCM clustering algorithm can be more effective and fast convergence, so as to improve the accuracy of urban traffic state identification.