Research on car-following model based on molecular dynamics

The car-following model has always been a research hot spot in the field of traffic flow theory. Modeling behavior can quantify longitudinal interaction between cars, thereby understanding characteristics flow, and revealing inherent mechanisms congestion other phenomena. In fact, there is an asymmetry problem driver’s acceleration deceleration operation. existing ignores difference cars. To solve this problem, cars driving on road are compared to molecules with interactions. Based molecular potential function wall function, we construct model. We use NGSIM data set calibrate parameters through genetic algorithm. Finally, analyze evolution rule disturbance different states help time-space diagram, compare classical optimal velocity results show that better describe from micro level.