Using a genetic algorithm for the design of an optimal transport network

Abstract

A transportation network serves the transport requirements of moving people and goods with different destination goals and relocation directions. The current network structure is usually a result of historically long adaptation process and the probability that it is not optimal is very high. Additionally it can be observed a growth of transportation needs. In these circumstances when a modernisation or expansion is required a number of competing designs must be evaluated. Combined total building expenses and maintenance costs are accepted among the evaluation criteria. Such a restriction does not guarantee an optimal solution as only a small fraction of the solution space is analysed. The input data for the optimisation problem cannot be entered in analytical form so it is natural to propose a genetic algorithm for performing the task.