Urban Road Network Design And Evaluation In Multimodel Transportation Network

Currently,traffic jams and multi-modal transportation are two problems which urban transport network needs to solve.This paper adds link capacity constraints to reflect the limitation of capacity on traffic volume,which contributes to describing the actual network state.A combined travel mode named Park-and-Ride is designed to reflect mixed traffic flow more accurately,which can also release traffic jams.Park-and-Ride refers to commuters use private cars to complete part travel which aims to arrive at transfer stations.After parking their private cars,commuters use public transport to arrive at destinations.Considering link capacity constraints,this paper constructs an equilibrium model about congested transport network.In this paper,KKT conditions are used to prove that optimal solutions satisfy UE assumption and uniqueness.Augmented Lagrange multiplier algorithm is used to solve congested network model,which combines Frank-Wolfe algorithm and external penalty function method.Logit model is adopted to divide commuters' travel modes aimed to study the impact of multi-modal transport network.Based on combined model,a multi-modal transport network equilibrium model with capacity constraints is built.Similarly,this paper proves that optimal solution of path variables satisfies UE assumption and traffic demand variables coincides with MNL model,which takes advantage of KKT condition.Optimal solutions' uniqueness is also verified.A method combining Augmented Lagrange multiplier algorithm and Evans algorithm is designed to deal with this multi-modal network equilibrium model.The chapter 3 of this paper studies a bi-level programming model of multi-modal network design problem with link capacity constraints.The upper level programming aims at the minimum sum of total network impedance and construction cost.This part closely relates to transport managers' benefits,and directly evaluates the rationality of network design.The lower level programming model is multi-modal network equilibrium model with capacity constraints,which is able to simulate commuters' path selection behaviors.In this paper,genetic algorithm is designed to handle this bi-level programming model about multi-modal network design problem with link capacity constraints.In chapter 4 of this paper,the feasibility of models which were established are verified.This paper respectively uses two cases to analyze models' each function blocks.Case 1 utilizes a small simple network for verification of models' functions by hand calculations.Case 2 checks a relatively complicated and larger Nguyen-Depuis network by C programming language.