Master of Science in Civil Engineering (MSCE)


Civil and Environmental Engineering

Document Type



In this study, alternative trip generation models for hurricane evacuation movement were developed using logistic regression and neural networks. The southwest Louisiana post-Andrew survey data were used for model estimation, validation, and comparison. The performance of the alternative models was compared with each other as well as against that of an existing evacuation model, the PBS&J model, developed for the same area. The results showed that the models developed in this study displayed similar performance. It was also found that the models developed in this study performed better than the existing PBS&J model in predicting household evacuation trip generation for southwestern Louisiana. The independent variables found to be significant in explaining household evacuation behavior included housing type, whether the household gets a mandatory evacuation order or not, age of the respondent, distance of the household from the closest body of water, and marital status of the respondent. Comparison of two model specifications involving different numbers of independent variables showed that the more comprehensive specification added very little to the explanatory power of the models and should be abandoned for model parsimony and ease of use.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Chester G. Wilmot