Resolution of an Antenna-Satellite assignment problem by means of Integer Linear Programming

Every day, ground stations need to manage numerous requests for allocation of antenna time slots by customers operating satellites. For multi-antenna, multi-site ground networks serving numerous satellite operators, oftentimes these requests yield conflicts, which arise when two or more satellites request overlapping time slots on the same antenna. Deconflicting is performed by moving passes to other antennas, shortening their duration, or canceling them, and has frequently been done manually. However, when many conflicts are present, deconflicting becomes a complex and time-consuming when done manually. We propose an automated tool that solves the problem by means of Integer Linear Programming. The models include operational constraints and mimic the manual process but consider the problem globally, thus being able to improve the quality of the solution. A simplified shortening model is also included to avoid excessive computation times, which is crucial given that the general problem has been reported NP-complete. Priorities are taken into account by tuning the cost function according to specifications of the requesting clients. Experiments with real-data scenarios using opensource software show that our tool is able to solve the Antenna-Satellite assignment problem for a large number of passes in a short amount of time, thus enormously improving manual scheduling operations, even when performed by a skilled operator. (C) 2014 Elsevier Masson SAS. All rights reserved.