1 An Integer Linear Programming model for Airline Use (in the context of CDM) Intro to Air Traffic Control Dr. Lance Sherry Ref: Exploiting the Opportunities of Collaborative Decision Making: A model and Efficient Solution Algorithm for Airline Use, Paul M. Carlson, Transportation Science, vol34, No. 4, Nov Briefing by: Babak. G. Jeddi Feb 2005
2 An IP model for Airline Use in the context of CDM A term Arrival Capacity: max # of possible landings without safety reduction Improving traffic flow mgt when weather condition reduces an airport’s Arrival Capacity During periods of reduced capacity (undersupply), FAA’s role shifts (should shift) from centralized DM to information gather and resource intermediary/negotiator Seeking Collaboration between FAA and Airlines CDM: A partially-decentralized Air Traffic Mgt (ATM) environment in which the airlines have increased decision making authority
3 The Problem Hubbing: the process of routing-origin destination traffic through a connecting airport rather than serving it non-stop! Requires careful scheduling to balance the conflicting goals of linking the greatest # of city pairs while minimizing the time that passengers spend at the hub. Many flights arrive within a short interval time (arrival bank), and many flights (departure bank) are scheduled to depart in a short time. Problem: schedules are very sensitive to disruption, due to the existence of connection dependencies. Hubbing airlines are often disrupted by weather and other factors! In 1996, disrupted flights caused loss of $440 million for Delta Airlines So the problem is very serious!
4 Solution! Question: Which flight should land first (prioritization)? A term Arrival slot: a time window during which the holder of the slot has permission to land one aircraft. FAA regulates the use of the scarce arrival capacity by allocating arrival slots among the airlines; Ground Delay Program, CDM, etc CDM has been replacing Ground Delay Program CDM: FAA informs the airlines about the state of the system and allocate constrained resources among the airlines; each airline will then decide/manage own operations according to its own priorities and objectives. FAA estimates that CDM would save about $2.6 billion over eight years (CDM homepage, 1997).
5 The airline must assign its inbound flights to its limited arrival slots, canceling or significantly delaying those flights without a slot. Also, decide whether to delay the completion times of inbound banks and by how much (called “spreading the bank”), which will determine how many of its bank flights will remain in their bank and how many will be separated. Using flight cancellation, flight separation, flight delay, and bank spread costs, the airlines seek the lowest cost solution among all the set of feasible solutions to the problem.
6 An OR (IP) model ! Objective function: – minimize the total cost of changes to scheduled operations, i.e. min sum of the costs of bank spread, flight delay, flight cancellation, and flight separation Constraints: – All banks should be eventually completed – All the flights in the bank to have either arrived, been canceled, or been separated – All flights are either arrived or cancelled – Arrival slot capacity restriction – …
7 A model for Airline Use in the context of CDM An Integer Linear Programming Model (by P. Carlson 2000): Decision variables:
8 A model for Airline Use in the context of CDM
9 Final point Model can be solved in real time (2 minutes)
10 Comments and Questions, please!