Impact of personalised route recommendation in the cooperation vehicle-infrastructure systems on the network traffic flow evolution

Inspired by the prevailing recommendation system application, personalised travel factors are introduced into route recommendation in order to provide more human-oriented travel service. With real-time information provided by the cooperation vehicle-infrastructure systems (CVIS), four real travel factors including distance, grade, time and toll are adopted to construct a route feature vector and an individual traveler preference feature vector, respectively. A novel route recommendation model based on Pearson’s correlation coefficient is formulated. A searching algorithm of all feasible routes is designed that achieves a better balance of time and space complexity. Considering that the traveler has heterogeneity in the numerous ways of using route recommendation information and choosing a satisfactory route, individual compliance with the route recommendation is creatively proposed and used to imitate a day-to-day route choice. A specific simulation with Monte Carlo method is conducted on a test network to show the dynamic evolution features of network traffic flow.