A GA-based transport modal choice model
This paper presents a hybrid transport modal choice model in which the genetic algorithm is applied for estimating parameters of a Multinomial Logit Model. The model has a simple decision structure which requires relatively modest computational capabilities and time to be estimated. Linear Utility Functions are defined for the transport modes which are included in the model structure as the fitness function. The objective function seeks to reduce the errors in the modal choice process. Moreover, the hybrid model incorporates distance as a spatial attribute variable for explaining commuters’ modal choices among automobile, bus, bicycle and walk modes. After this introduction section, a brief review on GA’s is presented. Section 3 introduces the Hybrid Transport Modal Choice Model structure, which is followed by the modeling results. Finally, Section 5 discusses this work’s findings and makes recommendations to further research.
