Title: A stochastic user equilibrium assignment problem in discrete network design problem

Authors: Masoomeh Divsalar; Reza Hassanzadeh; Iraj Mahdavi; Babak Shirazi

Addresses: Mazandaran University of Science and Technology, Babol, Iran ' Mazandaran University of Science and Technology, Babol, Iran ' Mazandaran University of Science and Technology, Babol, Iran ' Mazandaran University of Science and Technology, Babol, Iran

Abstract: We formulate the transportation discrete network design problem (DNDP) as a mixed-integer bi-level mathematical problem, based on the concept reserve capacity. The upper level goal programme maximises the reserve capacity by designing the direction of street and increasing the street capacity through lane addition. The lower level problem is stochastic user equilibrium traffic assignment problem within a probit-based path choice decision framework which generates user equilibrium flow patterns. Because of non-convexity nature of the model, meta-heuristic methods used to solve the problem and we used Monte Carlo simulation approach to compute path choice probabilities and stochastic user equilibrium solved by method of successive averages. A hybrid genetic algorithm with simulated annealing and an evolutionary simulated annealing algorithm are proposed. Numerical examples are presented to verify the proposed model and algorithm.

Keywords: discrete network design; reserve capacity; mixed-integer bi-level programming; stochastic user equilibrium; transport networks; street direction; street capacity; lane additions; Monte Carlo simulation; path choice probabilities; genetic algorithms; simulated annealing.

DOI: 10.1504/IJOR.2016.077682

International Journal of Operational Research, 2016 Vol.26 No.4, pp.422 - 442

Received: 27 Nov 2013
Accepted: 28 May 2014

Published online: 13 Jul 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article