星上传感器侧摆次数有限,在卫星观测调度时进行任务聚类可以节省资源,提高观测效率。提出了动态聚类调度算法(DCSA)来解决多星多轨道圈次的观测调度问题,DCSA把聚类和调度动态结合起来,并使用模拟退火算法搜索全局最优解;分析了在满足分辨率需求前提下多任务聚类的约束条件,并对聚类任务的侧摆角度和时间窗口进行合理优化;根据任务聚类的性质建立了聚类图模型,采用团划分方法进行DCSA中的聚类操作。多个实验算例表明,DCSA具有较好的求解性能,能有效提高卫星的观测效率。

Task clustering in satellite observation scheduling can save resources and improve observing efficiency. A dynamic clustering scheduling algorithm (DCSA) is proposed to solve the scheduling problem of multi-satellites and multi-orbits. DCSA merges the clustering operation and the scheduling process dynamically, and uses a simulated annealing algorithm to search for a global optimization solution. The constraints of multi-task clustering are investigated with the prerequisite of satisfying the resolution requirements of clients, and the slew angle and time window of cluster-tasks are optimized. A clustering graph model is constructed, and the clique partition method of graph theory is applied to task clustering operation in the procedure of DCSA. Computation results demonstrate that DCSA is feasible and can effectively improve satellite observing efficiency.