This study forecasts the maximum car ownership in a city that is consistent with environmental sustainability. A bi-level optimization model is used, where upper-level establishes car ownership consistent with the maximum environmental load on the road network, while the lower level assigns traffic demand across the network. Modal split and traffic environmental load models are used to connect the two levels. An algorithm, embracing sensitivity analysis (an acquiring derivative function of link flow and traffic demand with respect to zonal car ownership) is developed. To estimate the traffic environmental load accurately, an artificial neural network model is used to calculate the pollutants concentrations along the roads. (A) Reprinted with permission from Elsevier.
Abstract