Local Search and Genetic Algorithms for Satellite Scheduling Problems

Satellite scheduling is a family of problems that arise in satellite mission planning of communications of spacecrafts with ground stations. The problems of this family can be classified under time window scheduling given that the communication between the ground stations and spacecrafts can only be done during some specified time window. This later feature, together with the fact that different users may request communications with spacecrafts in the same time window, make the problems highly constrained and hard to solve to optimality. Therefore, these problems are solved in practice through heuristic and meta-heuristic approaches, which are in general efficient not only for small to moderate size but also for large size instances of the problem. While such methods do not provide any guarantee about optimality of solutions, in most cases they do provide high quality solutions that meet expected requirements of users. In this paper we consider the use of heuristic and meta-heuristic methods for solving satellite scheduling problems. We consider some local search methods (Hill Climbing, Simulated Annealing and Tabu Search) and population based methods (Genetic Algorithms and their variants). We show the instantiation of these methods for solving the case of Ground Station Scheduling, which is one of the most used variants of the satellite scheduling. Some computation results are presented for Tabu Search method.