Algorithm Research To Capacitated Traffic Assignment Problem

Compared with the user equilibrium model,models for the link capacitated traffic assignment problems(CTAP)can offer more realistic traffic assignment results.Previous studies of solution techniques for the link capacitated traffic assignment problems are summarized in this paper.On the basis of the review,the target and content of this paper are proposed.Models for the link capacitated traffic assignment problems can be a point of penetration.The karush-kuhn-tucker conditions are the analysis tools for equilibrium conditions of CTAP.Based on analysis of solving process for gradient projection,The principles and converge property of inner penalty and augmented lagrangian multiplier are discussed and implemented.NguyenDupuis network is used for numerical testsNext,based on the equilibrium conditions of CTAP,characteristics of oversaturated links and solving process for inner penalty or augmented lagrangian multiplier,link performance functions is established.Then,the paper constructs model of CTAP and proves the existence and uniqueness of traffic assignment results.Combined update features of road delay items and gradient projection,the Heuristic approach of CTAP is obtained.Nguyen-Dupuis network is used for numerical tests.Compared with the classical solution algorithm,the heuristic algorithm converges faster with the same convergence accuracy.Finally,Sioux-Falls network and kunshan network are used for numerical tests,Beckmann objective function value,volume per capacity ratio,residual flow on congestion section and delay will be the indexes that verifies the applicability of the algorithm for large road network.The results show that the heuristic algorithm has good applicability to large-scale road networks.