Research Of Commuter Route Choice Behavior And Coordinative Control Optimization Based On License Plate Data

The rapid growth of motor vehicle ownership and the boost of economic make the urban traffic system face great pressure and the traffic congestion problem is becoming serious.As the rigid demand of most travelers,car-based commuter travel is the key factor affecting the quality of urban traffic during the peak period.Therefore,the travel information mining,the study of travel rules and the corresponding management and control strategy optimization are of great significance to alleviate the traffic congestion during the rush hour.This paper first establishes a method of commuting trajectory extraction based on license plate data.K-means++clustering is used to identify vehicles with commuting characteristics in the traffic network,and the longest common sub-sequence(LCSS)method considering time constraints is proposed to calculate the similarity of different trajectories,so as to reconstruct the space-time commuting trajectory according to individual commuting trajectory similarity,travel time matching degree,road topology information and so on.In order to deeply reveal the route choice behavior of commuter individuals,this paper further analyzes the traffic characteristics of the actual chosen path and the optimal solution(shortest travel time path,shortest distance path)in commuting travel,the distribution of the path length,travel time,detour degree,intersection density of the three kinds of paths is compared,and the coincidence between the actual choice path and the optimal scheme is analyzed.On this basis,the Path-Size Logit based commuting path selection model is established combined with the alternative path set search method based on dynamic K-shortest path.In addition,the stability of departure time and path selection of commuters are analyzed in combination with specific cases.Combining the travel choice characteristics of car commuters,this paper establishes a bi-level programming model of coordinative optimization of signal control and traffic guidance.The upper layer of the model is signal control optimization aiming at the minimum total delay of all vehicles at intersections,and the lower one is commuting traffic flow allocation optimization aiming at the shortest total travel time of vehicles in road network.At the same time,a probabilistic model of commuting path conversion is established to obtain the diverted commuter flow of different road segments.The genetic algorithm and the Frank-Wolfe method are used to solve the upper and lower models respectively,and the optimization results are input and iterated with each other until the overall optimization requirements of the bi-level programming are met.Based on the case analysis of Jinan local road network,this paper verifies the effectiveness of the commuting trajectory reconstruction algorithm and the commuting path selection model.Finally,a case of road network in Jinan is analyzed to verify the effectiveness of the commuter trajectory reconstruction algorithm and commuter route choice model;on this basis,the coordinative optimization of signal control and commuter traffic assignment is applied in a small-scale road network.Experimental results show that the coordinative optimization for commuter vehicles is superior to either single signal control or traffic allocation optimization.