Bilevel optimization of a housing allocation and traffic emission problem in a predictive dynamic continuum transportation system

Abstract

In recent decades, the effects of vehicle emissions on urban environments have raised increasing concerns, and it has been recognized that vehicle emissions affect peoples’ choice of housing location. Additionally, housing allocation patterns determine people's travel behavior and thus affect vehicle emissions. This study considers the housing allocation problem by incorporating vehicle emissions in a city with a single central business district (CBD) into a bilevel optimization model. In the lower level subprogram, under a fixed housing allocation, a predictive dynamic continuum user‐optimal (PDUO‐C) model with a combined departure time and route choice is used to study the city's traffic flow. In the upper level subprogram, the health cost is defined and minimized to identify the optimal allocation of additional housing units to update the housing allocation. A simulated annealing algorithm is used to solve the housing allocation problem. The results show that the distribution of additional housing locations is dependent on the distance and direction from the CBD. Sensitivity analyses demonstrate the influences of various factors (e.g., budget and cost of housing supply) on the optimized health cost and travel demand pattern.