Integration of selecting and scheduling urban road construction projects as a time-dependent discrete network design problem

• Three different bi-level programming models based on DNDP are developed.
• A general framework and an evolutionary approach are proposed to solve the model.
• A procedure is proposed to reduce significantly the model solution time.
• Computational tests show acceptable performance of the proposed approach.

Decision making on the selection of transportation infrastructure projects is an interesting subject to both transportation authorities and researchers. Due to resource limitations, the selected projects should then be scheduled during the planning horizon. Integration of selecting and scheduling projects into a single model increases the accuracy of results; however it leads to more complexity. In this paper, first, three different mathematical programming models are presented to integrate selecting and scheduling of urban road construction projects as a time-dependent discrete network design problem. Then, the model that seems more flexible and realistic is selected and an evolutionary approach is proposed to solve it. The proposed approach is a combination of three well-known techniques: the phase-I of the two-phase simplex method, Frank-Wolfe algorithm, and genetic algorithm. Taguchi method is used to optimize the genetic algorithm parameters. The main difficulty in solving the model is due to the large number of subsequent network traffic assignment problems that should be solved which makes the solution process very time-consuming. Therefore, a procedure is proposed to overcome this difficulty by significantly reducing the traffic assignment problem solution time. In order to verify the performance of the proposed approach, 27 randomly generated test problems of different scales are applied to Sioux Falls urban transportation network. The proposed approach and full enumeration method are used to solve the problems. Numerical results show that the proposed approach has an acceptable performance in terms of both solution quality and solution time.