Human Behavior and Optimization
Choice models have been the workhorse of travel demand modelers in transportation since the 1970s. They are designed to capture complex aspects of human behavior and are operational in the sense that they are used intensively by practitioners. Still, they have not made their way to the vast field of operations research, or only in a timid fashion.
In this lecture, held by Michel Bierlaire (Ecole Polytechnique Fédérale de Lausanne), we will try to link choice models and optimization.
First, we will show that the first principles of choice modeling are actually founded on optimization. We will then give a short introduction to some of the most popular choice models. Using a linear specification, we will then discuss how they can be introduced into non-trivial Mixed Integer Linear Problems. Finally, we will show how scheduling algorithms can be used to model complex travel behavior.