A Comparison of Iterative Repair Strategies for Earth Observing Satellites Imaging Scheduling

Earth Observing Satellites imaging scheduling problem needs to assign time slot of the participated satellites to the tasks, the object of which is to maximize the total evaluation of the assigned tasks without constraint violation, thus is an oversubscribed scheduling problem. As iterative repair method are successfully used in the oversubscribed scheduling problem, we propose an iterative repair based heuristic method to solve it, which is inspired from Kramerpsilas AMC scheduling problem. We redesign the task selection and re-arranging procedure, and try some heuristic methods to test the performance. These heuristic methods include time-sequence based search, photo-probability based search and stochastic search. Finally, we compare these approaches on two types of data-sets. The result shows that, photo-probability based search performs best when large number of tasks are participated, while with less tasks, the stochastic search performs best, time-sequence based search is almost always outperformed by photo-probability based search.