Alleviating Travel Delay Uncertainties in Traffic Assignment and Traffic Equilibrium

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Abstract

This grant provides funding in order to tackle traffic equilibrium, traffic assignment and route guidance problems when travel costs/delays are uncertain. The main goals are to alleviate congestion and better manage traffic in the face of ?uncertain conditions?. Our goal is to utilize ideas from three fields: stochastic optimization/ stochastic variational inequalities, robust (and adjustable robust) optimization and finally, learning theory from Statistics. These approaches will allow us to address the aspect of uncertainty in data parameters in a tractable way. Our research contributions will be to (i) design and study a variety of formulations - involving stochastic optimization, robust optimization and learning - of key traffic flow planning and management models, (ii) understand how to incorporate effectively the inherent element of stochasticity in the nature of traffic delays in a way that does not depend on distributional assumptions on the data, (iii) examine efficient solution methods of these problems and finally, and (iv) validate our results computationally. If successful, the results of this project will provide solutions to key traffic planning and management problems that are robust to uncertainty in terms of travel delays. The potential applications and hence impact of this research to a variety of logistical networks are very significant. In the area of transportation, this research will impact both the fields of Advanced Traveler Management Systems (ATMS) and Advanced Traveler Information Systems (ATIS). This research aims at using methodologies from a variety of fields with the long term goal to not only advance knowledge and deepen our understanding of issues in transportation and as they relate to ATMS and ATIS, but also to contribute to these fields by developing creative and innovative concepts. The goal is to build an integrated framework, models and solution techniques for the application of stochastic optimization and robust optimization to key traffic planning and management problems and related areas that are accessible to all, in order to help transportation academics and practitioners. From an educational perspective, the results of this project will serve as components in teaching modules at MIT. These include modules in core courses the PI has already been one of the key contributors. This project lends itself ideally to mentoring undergraduate and graduate students.