019710 – Models for demand analysis

1. Course Objectives
In this course we will provide the tools necessary for analysis and forecasting of the demand for services and goods. It will address the theory of demand and prediction of demand with a focus on discrete choice models and reliance on statistical approaches and behavioral theories. We will also address innovative approaches to demand analysis using various examples.
Applications will revolve around the demand for travel in transportation systems and the role of the models that we will study in the planning of urban and regional transportation systems. We will use examples of advanced models that have been developed in various applications.

2. Course Content
The course will comprise of two main parts. The first part of the course will focus on the theory of discrete choice models. The second part will be devoted to the specification and estimation of these models. Lecture subjects will be as follows:
• Introduction – requirements of transportation demand models , new trends in transportation planning , aggregate and disaggregate behavioral models
• Statistical estimation methods – linear regression, likelihood function , properties of estimators, statistical tests
• Model specification – definition of the model structure, specification of utility functions, dummy variables, nonlinear specification.
• Discrete choice theory – deterministic choice models, stochastic choice models, assumptions of the Logit and Probit models.
• Binary Logit model – Definition and properties, estimation methods
• Multinomial Logit model – Definition and properties, limitations, and the IIA property
• Nested Logit model – Definition and properties, estimation methods,
• Construction of transportation demand models – the model development process, practical difficulties in estimation
• Statistical tests for discrete choice models
• Aggregation and forecasting – Aggregation of nonlinear models, prediction with discrete choice models
• Experimental design, stated and revealed preference data,
• Model systems, advanced models and current trends in model development.

3. Bibliography
The course is based on the book: Moshe Ben- Akiva & Steven Lerman ( 1985 ) Discrete Choice Analysis : Theory and Application to Travel Demand, MIT PRESS
Additional readings will be distributed during the course
4. Course structure and grading
The course will consist of weekly lectures, assignments (50% of the grade) and a final exam (50% of the grade).