Abstract
This study attempts to develop a macro-level car ownership model using a bi-level optimization modeling approach. The upper level of the bi-level model deals with a maximum problem of zonal car ownership. Objective function is the total zonal car ownership and the constraints are the legalized standard of air quality and the frontier emissions estimated using a stochastic frontier analysis approach. The lower level is a user equilibrium assignment model. Pollutant concentrations are estimated using an artificial neural network model. The interdependencies of car ownership, traffic flow, and the emissions and pollutant concentrations are logically represented based on an iterated optimization process. The final optimized car ownership can be used as a benchmark of realizing environmentally sustainable transportation systems. Based on the data collected in Dalian, China and the Millennium Cities Database, the effectiveness of the proposed car ownership model was empirically confirmed.
