Comparing Meta-Heuristic Algorithms for Solving Transportation Network Design Problem

Please enable javascript in your browser.

Farivar, Saeedeh | 2009

743 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 39520 (09)
University: Sharif University of Technology
Department: Civil Engineering
Advisor(s): Zokaei Aashtiani, Hedayat
Abstract:
Discrete network design problem is selecting a set of feasible projects among a set of candidate projects, considering a budget constraint in which user's cost is being minimized. Network design problem is one of the complicated problems in transportation science, since it is a bi-level problem in which the lower level is a traffic assignment problem and the upper level is finding the best set of arcs. Meta-heuristic algorithms can be used to solve this problem. Although, there is no guarantee to find the optimal solution by Meta-heuristic algorithms, but they find good solutions in short times. The purpose of this study is comparing the efficiency of three Meta-heuristic algorithms for solving transportation network design problem, including Genetic, Ant System, and Simulated Annealing algorithms. The results of applying these algorithms for Sioux Falls network show that the Ant System Algorithm has the best performance in terms of both number of traffic assignments and the accuracy of the optimal solution. Afterwards, the Simulated Annealing Algorithm has an acceptable performance, while the Genetic Algorithm has the poorest performance.
Keywords:
Genetic Algorithm ; Ant System Algorithm ; Meta Heuristic Algorithm ; Simulated Annealing Method ; Discrete Network Design problem

No TOC