Scheduling a constellation of agile earth observation satellites with preemption

Abstract

In this paper, we consider a scheduling problem for a set of agile Earth observation satellites for<br><br>scanning different parts of the Earth’s surface. We assume that preemption is allowed to prevent repetitive<br><br>images and develop four different preemption policies. Scheduling is done for the imaging time window and<br><br>transmission time domain to the Earth stations as well. The value of each picture from different target<br><br>regions and the limitations of the satellite constellation in terms of memory and energy cause high<br><br>computational complexity for this problem and thus obtaining an optimum solution with a deterministic<br><br>method is very time-consuming. Consequently, a genetic-based metaheuristic algorithm with a specific<br><br>solution representation is developed in order to maximize the total value of the observation process by<br><br>establishing heuristic rules in the initial population of this algorithm. Comparison of the results from the<br><br>proposed model with the results of cases where repetition of observed areas is not ignored indicates that the<br><br>proposed model can bring about a significant increase in profits in the planning horizon.