Tail Assignment Optimization

· Reducing fuel consumption

· Producing feasible flight schedule for operation

· Reduce business unit efforts for replanning

Tail Assignment Optimization is now in production at Turkish Airlines for some sub-types, and expanding process to other sub-types is going on. Preliminary results show that Tail Assignment Optimization can deliver schedules which has less fuel in fuel consumption, up to 0.2%, at the beginning of the operational period, which equates to 12 million USD annually for Turkish Airlines.

Due to the performance and benefits of the software, it has been adopted by flight operations teams, and a continuation project has been requested for the addition of new features. Additionally, negotiations have started for sales to different companies, and POC processes are being initiated.

One server, a development platform, open source mathematical solver

Implementation Complexity

Implementation complexity of Tail Assignment Optimization could be analyzed in two parts:

Integration Complexity:

Input integration: There are different data types used in Tail Assignment Optimization. Flight data, maintenance data, constraint data, fuel performance data, fuel price data, ground time data and fuel consumption data. These data are gathered from different sources via ETL, stream topics and web services and written to a database. Tail Assignment Optimization connects to this database and read required data from here.

Output integration: There are two integration points in output side. First one is the database integration. Results of the optimization process are written to database. Second one is the middleware connection. Tail Assignment Optimization triggers middleware to send the output data from database to operation management software.

Development Complexity:

Tail Assignment Optimization utilize different methods to solve this complex problem. There are algorithms for analyzing, modifiying the data and to conduct preprocess. Then a novel mathematical model is designed on top of these algorithms. This methodology requires a scientific backround and causes a high level of implementation copmplexity.

One of the most challenging problems in the aviation sector is determining which flights will be performed by which aircraft. During the scheduling period, it is decided by the scheduling teams which sub-type aircraft will operate on which flight. Thus, while it is known which sub-type will execute the flights up until the operational period (the last three days), it is not specifically known which aircraft will perform them. The assignment of specific aircraft to flights is done when transitioning from the scheduling period to the operational period, and this process is called the tail assignment process.

At Turkish Airlines, the tail assignment is carried out by a module of a software which is used by the company to manage operational processes. However, these assignments are done at a basic level without evaluating fuel consumption or flight-aircraft constraints.

The Tail Assignment Optimization Project emerged from the need to minimize fuel consumption and create feasible schedules. Even among same sub-type, there can be differences in fuel performance of aircrafts. This leads to variations in fuel consumption based on the assignment decisions. On the other hand, due to equipment or authorization constraints on the aircraft, not every aircraft can be assigned to every flight, and correcting these infeasible assignments can be a tedious process.

The Tail Assignment Optimization software takes these inputs and constraints into account to produce a schedule that minimizes fuel consumption and is feasible in terms of flight-aircraft constraints. In this software, sophisticated solution methods are utilized, primarily mathematical programming, which is highly effective and scientific in solving complex problems. Hence, Tail Assignment Optimization can provide a safe and effective solution in a short time.