Bi-objective bimodal urban road network design using hybrid metaheuristics

In this paper a bimodal discrete urban road network design problem with bus and car modes is investigated. The problem consists of decision making for lane addition to the existing streets, new street constructions, converting some two-way streets to one-way streets, lane allocation for two-way streets, and the allocation of some street lanes for exclusive bus lanes. Two objectives are considered in the problem: maximization of consumer surplus, and maximization of the demand share of the bus mode. The interaction of automobile and bus flows are explicitly taken into account and a modal-split/assignment model is used to obtain the automobile and bus flows in the deterministic user equilibrium state. The main contribution of the paper lies in proposing a new network design problem that combines the road network design decisions with the decision making for bus networks. The problem is formulated as a mathematical program with equilibrium constraints. A hybrid of genetic algorithm and simulated annealing, a hybrid of particle swarm optimization and simulated annealing, and a hybrid of harmony search and simulated annealing are proposed to solve the problem. Computational results for a number of test networks are presented and investigated.