Time-dependent transportation network design that considers health cost

This paper proposes a bi-level optimisation framework for time-dependent discrete road network design that considers health impacts. A general health cost function is proposed and captured in the framework. The function simultaneously considers the health impacts of road traffic emissions, noise, and accidents due to network expansion. To solve the problem, the artificial bee colony (ABC) algorithm is proposed to search the network design solutions of the upper-level problem, while the method of successive averages and the Frank-Wolfe algorithm are adopted to solve the lower-level time-dependent land-use transportation problem. A repairing procedure is proposed to remedy infeasible solutions. A numerical study is set up to illustrate the conflict between maximising consumer surplus and minimising the health cost. This paper also reveals a paradox phenomenon that with an increasing amount of emissions, the health cost decreases. Moreover, the existence of a health inequity between different residential zones is demonstrated. A modified Sioux Falls network is adopted to show the performance of the solution algorithm as well as the effectiveness of the proposed repairing procedure.