Leader-follower green traffic assignment problem with online supervised machine learning solution approach

Abstract In this paper, we propose a bi-level traffic assignment problem where at the upper-level the total traffic CO emission objective function is evaluated and at the lower level, the traffic assignment objective function is considered. At the upper level, the total emission objective function is considered as the main objective from a macroscopic view point of the system manager. At the lower level, the traffic assignment problem is considered to optimize the users’ travel times, individually. Although the lower-level objective function is convex with some linear constraints, the proposed bi-objective problem is np-hard, and even finding a near-optimal solution is also np-hard. Hence, we applied a pioneer online supervised machine learning (SML) algorithm to find the optimal solutions in a reasonable running time. The validity of the online SML is verified through some real urban transportation examples in medium and large-sized networks.