The Bi-level Programming Model And Algorithm For Discrete Network Design Problem Under Factors Affecting Exhaust Emission

With the acceleration of urban motorization and increasing attention to environmentalissues in our country, the study of traffic network design problem under the factors affectingexhaust emission not only has important theoretical value, but has practical significance insolving environmental and traffic problems.This paper analyzes the research status of network design problem and its relevanttheoretical knowledge firstly, then builds a bi-level programming model about the discretenetwork design problem (DNDP) under the factors affecting exhaust emission, and designs analgorithm to solve the model. The main contents are studied as follows:(1) Building a DNDP model under the factors affecting exhaust emission based on thebi-level programming thesis, in which the upper level is the goalpp of CO emissions andsystem total travel time on condition that vehicles travel in uniform motion, the lower level isa standard user equilibrium assignment.(2) The vehicle emission model, travel cost and traveler behavior are studied. Designinga hybrid algorithm based on genetic algorithm and user equilibrium assignment algorithm.(3) Taking the classic Sioux Falls network as example for solving the model in differentcondition. On the one hand, on the basic of fixed investment budget and mutation probability,analyzing the influence population size and crossover probability makes on objective solution,determining optimum population size and crossover probability, and obtaining the trafficnetwork design scheme under different weight. On the other hand, in order to maximize theinvestment benefit, calculating and analyzing the traffic network design scheme in differentbudget level on the foundation of the obtained optimum population size and crossoverprobability.