Using a genetic algorithm to optimize transportation network in Hong Kong

Using a genetic algorithm to optimize transportation network in Hong Kong

0.00 Avg rating0 Votes
Article ID: iaor2002841
Country: China
Volume: 20
Issue: 7
Start Page Number: 94
End Page Number: 98
Publication Date: Jul 2000
Journal: Systems Engineering Theory & Practice
Authors: ,
Keywords: genetic algorithms
Abstract:

Optimization of highway and public transport networks becomes important, especially in conditions of heavy concentration of urban population such as in Hong Kong. With the growth of population, better transport infrastructures are needed. Some three billion dollars have been invested every year for the construction of new transport infrastructures in Hong Kong. These new transport infrastructures have been chosen from a large number of potential highway and public transport projects. This problem can be formulated as a 0–1 programming problem mathematically. In this paper, a genetic algorithm is used to solve the 0–1 programming problem in which some new highways and railway infrastructures are identified as the potential projects in Hong Kong so as to minimize the total development and transportation cost. It is the first work to apply genetic algorithm for network design in practice. Three cases are considered using genetic algorithm to optimize the highway and public transport networks. The first one is to optimize the highway links which are selected from the 254 potential highway projects, while the second one is to optimize the public transport links which are selected from the 75 potential public transport projects. Finally, both the highway and public transport links are optimized simultaneously. The sets of solution are analysed. The developed model will advance this subject and help the authorities decide which new highways and public transport infrastructures should be considered for detailed investigation. The Hong Kong 2006 planning data and transportation network are used.

Reviews

Required fields are marked *. Your email address will not be published.