Generating Sustainable Solutions for the Urban Transportation Considering Multiple User Classes

Sustainability of transportation has become one of the most significant issues in the world due to the increasing global warming concerns. There has been a growing interest in improving the performance of the transportation systems using sustainability concepts. In this study, we use some social and environmental measures to assess the sustainability of transportation, such as accessibility of zones and equal distribution of the accessibility, accidents, emissions and noise generated by vehicles. Our main aim is to achieve a sustainable traffic assignment. To carry out our purpose, we develop a bilevel traffic assignment model. The lower level of this model consists of the stochastic user equilibrium with multinomial logit discrete choice and multiple user classes. At the upper level, we identify user class based toll prices while optimizing social and environmental objectives. As the two objectives are conflicting, we aim to generate a set of Pareto optimal solutions. Overall problem is solved with an adapted genetic algorithm, namely Non-Dominated Sorting Genetic Algorithm. During the course of the algorithm, the lower level problem is solved many times by means of Self-Regulated Averaging method. The developed method is applied to a known instance from the literature, and solutions are elaborated. Keywords—Sustainable traffic assignment, bi-level programming, stochastic user equilibrium, non-dominated sorting genetic algorithm, self-regulated averaging method