摘要

Disaster and environment monitoring need large amounts of in time and effective images. Agile earth observing satellite has connatural advantage to get these images. The character of agile satellites mission scheduling for disaster and environment monitor problem is long time windows and multi time windows constraints. The problem is divided into two sub-problems: task & resource matching problem and single satellite task arrange problem. A novel algorithm which separates the single satellite arranging from the intelligent optimization process is proposed solve the problem. A learnable genetic algorithm is proposed to solve the task & resource matching problem, selects satellite, time windows and task observing order for tasks; single satellite task arrange method arranges tasks and chooses task observing start time based on backward time slack. The knowledge model learns from satellites scheduling results, and guides the search process of genetic algorithm. Experiment results show that the effectiveness of the proposed approach.