Machine learning methods for transportation under uncertainty

Motivated by the prevalence of uncertainty and the widespread use of modeling in Transportation, we develop and study effective methods for modeling Transportation under uncertainty. These methods are Machine Learning-based, i.e., they extract patterns from data and leverage them for better modeling. We study them through several case studies, including: quick adaptation of traffic models upon road incidents; estimation of mobility demand from limited observations; and predictive optimization of dynamic Public Transport. Our results yield several positive conclusions about the effectiveness of the studied methods for current and future Transportation.Doctor of Philosoph