Events Calendar

Epstein Institute / ISE 651 Seminar Series

Speaker: Guglielmo Lulli, Assistant Professor, Dept of Informatics, Systems and Communication, University of Milano – Bicocca, Italy

Talk Title: " 'Facets' of the Air Traffic Flow Management Problem"

Series: Epstein Institute Seminar Series

Abstract: Air traffic has experienced a dramatic increase in recent years. During time periods when the weather is less than ideal and/or demand is at a peak, severe congestion is typically experienced. In response, the air traffic management (ATM) systems assign large delays to many flights. The resulting additional costs are very substantial. It is not surprising that the minimisation of delays due to congestion in air traffic networks has received a lot of attention both from aviation authorities and from the scientific research community. The most effective way to deal with the situation, at least in the short-term, is to adjust the flow of air traffic on a continuous basis so that it matches the available capacity of the various components of the ATM network, thus reducing flight delays. This is known as the Air Traffic Flow Management Problem (ATFM).

In the past few years, many mathematical models have been developed for this purpose, especially to address capacity constraints at major airports or in the terminal airspace around them. However, en-route capacity constraints also impose important limits on air traffic flows. Some en-route sectors experience congestion on an almost routine basis, especially during the summer months. Dealing with sector capacity restrictions requires new mathematical models that capture the largest possible set of feasible actions. In this new framework, the spectrum of available options is wider and it includes in addition to the traditional ground and airborne holding delays, both more tactical actions, e.g., "metering" and "miles-in-tail", and rerouting, i.e., the possibility of rerouting flights through alternative flight paths.

In this talk, we present a new integer programming formulation which captures all the aspects described above. The scope of the model is to suggest the time of departure, the route, the time required to cross each sector and the time of arrivals taking into account the capacity of en-route sectors and airports. The main feature of the model is the formulation of rerouting decisions in a very compact way. With respect to previous models, the methodology we presented does not require any additional variables, but it only introduces new constraints. These constraints implement local routing conditions that are sufficient for the purpose of the model. Moreover, to strengthen the polyhedral structure of the underlying relaxation, we also present three classes of valid inequalities. Several polyhedral insights are also provided.

A wide computational analysis on realistic instances demonstrated the viability of the proposed model. We report short computational times (less than 15 minutes) on instances of the size of the US air traffic control system that make it realistic that our approach can be used as the main engine of managing air traffic in the US. Given that our approach includes all the air traffic control decisions (ground holding, air holding, adjusting speed of aircraft and rerouting) combined with the attractive computational times, makes us optimistic that this approach may succeed in becoming the main air traffic control engine.

This is joint work with D. Bertsimas and A. Odoni.

TUESDAY, OCTOBER 15, 2013
GRACE FORD SALVATORI (GFS) ROOM 101
3:30 - 4:50 PM

Biography: Guglielmo Lulli is Assistant Professor of Operations Research at University of Milano-Bicocca. He received a PhD in Operations Research in 2003 from the University of Rome “La Sapienza”. During his studies, he visited the NEXTOR center at University of Maryland and the System and Industrial Engineering Department at University of Arizona, both for one year appointment. In 2007, he was recipient of a Fulbright Fellowship at Massachusetts Institute of Technology. His research interests focus on mathematical programming and stochastic programming particularly as applied to transportation and logistic operations, air traffic flow management and bio-computational problems.

Host: Daniel J. Epstein Department of Industrial and Systems Engineering

More Information: Seminar-Lulli.doc

Location: Grace Ford Salvatori Hall Of Letters, Arts & Sciences (GFS) - Room 101

Audiences: Everyone Is Invited

Contact: Georgia Lum