Earth observing satellites (EOS) orbit around the earth to perform observation tasks specified by users. The additional maneuverability resulting from higher degrees of freedom than nonagile EOS (N-AEOS) provides agile EOS (AEOS) a significantly larger visible time window to complete the tasks. As a consequence, the task scheduling for AEOS is much more computationally complex than N-AEOS. In this article, a mixed-integer nonlinear optimization problem is formulated to find a near-optimal task allocation for a realistic AEOS scheduling problem. The satellite resources, such as energy and memory constraints, are considered in this problem. A reward factor is used to address the requirement of multiple scans in order to complete a task. A probability factor is also taken into consideration to incorporate the uncertainty of successful scans due to external factors, such as cloud coverage. An elitist mixed coded genetic algorithm-based satellite scheduling (EMCGA-SS) algorithm is proposed to solve the formulated problem. EMCGA-SS is extended to elitist mixed coded hybrid genetic algorithm-based satellite scheduling by combining a hill-climber mechanism in order to have better initialization. Experimental results to illustrate the performance of the algorithms and a comparison with some widely used methodologies are also presented.

中文翻译：

---

具有能量和内存约束的基于奖励因素的多敏捷卫星调度


地球观测卫星（EOS）围绕地球运行，执行用户指定的观测任务。比非敏捷 EOS (N-AEOS) 更高的自由度带来的额外可操作性为敏捷 EOS (AEOS) 提供了明显更大的可见时间窗口来完成任务。因此，AEOS 的任务调度在计算上比 N-AEOS 复杂得多。在本文中，我们制定了一个混合整数非线性优化问题，为实际的 AEOS 调度问题找到接近最优的任务分配。该问题考虑了卫星资源，例如能量和内存限制。奖励因素用于满足多次扫描的要求以完成任务。还考虑了概率因素，以纳入由于云覆盖等外部因素导致的成功扫描的不确定性。提出了一种基于精英混合编码遗传算法的卫星调度（EMCGA-SS）算法来解决该问题。通过结合爬山机制，将 EMCGA-SS 扩展到基于精英混合编码混合遗传算法的卫星调度，以实现更好的初始化。还提供了说明算法性能的实验结果以及与一些广泛使用的方法的比较。