Equilibrium Analysis Of Urban Mixed Traffic Network Under The Influence Of Car Sharing

With the development of economy and the improvement of people's living standard,people's traveling demand and distance have been increasing which leads to a sharp increase of vehicle inventory.A large number of vehicles have resulted in a series problems such as traffic congestion,energy shortage and environmental pollution.In this case,‘Car Sharing' has emerged as the times require.At the moment,most of the research about car sharing is qualitative.There are a few quantitative research but they mainly focus on discussing the feasibility of introducing car sharing and the factors that affect residents choosing car sharing through empirical analysis.The thesis employs ‘Car Sharing' as an important transportation mode into multi-mode traffic networks and studies the equilibrium problem of mixed traffic network.Then,we study traffic congestion at the cross road.First of all,a mixed-mode traffic equilibrium model is constructed.Based on summarizing the previous research,a new time-flow impedance function which is more realistic is presented.Next,a mixed-mode network equilibrium model is constructed employing ‘Car Sharing' as a new traffic mode according to the new impedance function,in which battery electric vehicles and gasoline vehicles are both able to be chosen.Later,we suggest two approaches to handle the model: one is to use the interior point algorithm,and the other is to employ the sequential quadratic programming algorithm.Three numerical examples are presented to illustrate the efficiency of these approaches and the influence of distance limited and unit operating cost on the traffic mode and path selection.Then,a linear program model with linear complementarity constraints is presented.This part presents a linear program model with linear complementarity constraints to solve traffic signal optimization problem employing a weight variable to show the importance of the lane phases.The objective function of the model is to obtain the minimization of total queue length with weight factors at the end of each cycle.Next,an algorithm based on SQP algorithm is proposed,by which the local optimal solution can be obtained.Further,a numerical experiment is proposed to study how to set the initial solution of the algorithm that can get a better local optimal solution more quickly.