Scheduling Satellite Timetables using DCOP

Earth observation satellites (EOS) are satellites equipped with optical sensors that orbit the Earth to take photographs of specific areas at the request of users. With the development of space technology, the number of satellites increases continuously. Yet still, the number of satellites cannot meet the explosive growth of applications. Thus, scheduling solutions are required to satisfy requests and obtain a high observation efficiency. While the literature on multi-satellite scheduling is rich, most of the solutions are centralized algorithms. However, due to their cost, EOS systems are often co-funded by several agents (e.g., countries, companies, or research institutes) and central solutions require that these agents will share their requests for observations with others. To date, there is no solution for EOS scheduling that protects the private information of the interested parties. In this study, we model the EOS scheduling problem as a distributed constraint optimization problem (DCOP). This modeling enables generating timetables for the satellites in a distributed manner without a priori sharing private information of the users with some central authority. For solving the resulting DCOP, we use the Distributed Stochastic Algorithm (DSA), which is a simple DCOP algorithm that is known to produce efficient solutions in a timely manner. The modeling together with the solving of the resulting DCOP constitute our new solution method, which we term Distributed Satellite Timetable Solver (DSTS). Experimental evaluation reveals that the DSTS method provides solutions of higher quality than a commonly-used Greedy algorithm.