Bi-objective methods for road network problems with disruptions and connecting requirements

Abstract

Disruptions in urban roads can significantly alter the quality of the transportation network by generating more congestion, gas emission, noise, stress, etc. In some situations, it can even break the path between some pairs of nodes in the road network (strong connectivity in graph theory). To avoid this issue, traffic managers can temporarily change the orientation of some streets (arc reversal). In this study, we propose bi-objective methods for solving the bi-objective Unidirectional and bi-objective Multidirectional Road Network problems with Disruptions and connecting requirements (resp. bi-URND and bi-MRND). In bi-URND, the road network represents local networks such as city centers with narrow streets. In this case, a simple graph is used to model the transportation network. A more general urban network is addressed with bi-MRND by means of a multi-graph model. We propose an -constraint method to compute the Pareto-optimal fronts, using an up-to-date mathematical formulation and an NSGA-II. Both bi-objective methods are compared with two metaheuristics (a Biased Random Key Genetic Algorithm and an Iterated Local Search) proposed by Huang, Santos, and Duhamel (2019) and, including an aggregation of the two objective functions. Results are presented for simulated and realistic instances on Troyes city in France.