综合点目标和区域目标的多星观测调度

Multi-Satellite Scheduling Toward Spot and Polygon Observing Requests

卫星遥感应用中存在点和区域两类任务,两类任务的调度方式存在区别,将二者综合调度困难。本文将点目标视为特殊的区域目标,将它们按照卫星观测机会分解成元任务,并将两类任务统一成元任务。根据任务类型分别构建收益函数,兼顾了二者在收益计算上的差异,从而建立综合模型对两类任务统一处理。根据问题特征,设计了具有多重分化机制的模拟退火算法进行求解,通过随机扰动、重排列及重启动三种分化机制,避免算法陷入局部最优,以得到更优解。实例验证了方法的有效性。

Remote sensing requests can be divided into two types： spot targets and polygon targets. Satellite observation scheduling toward these two kinds of targets are different, which makes it hard to schedule the two kinds of requests synthetically. In this article, an integrated model was proposed to deal with two kinds of targets. The spot target is regarded as a special polygon and can be divided into subtasks. The subtasks are generated according to the observing chance of the target, and be mapped into atomic task. Two evaluation functions are put forward to evaluate the profits of different targets. A muhi-diversifications simulate annealing algorithm is proposed for the satellite observation scheduling problem. Three diversification mechanisms： perturb, rearrange and restart, are defined to avoid the local optimum solutions, enlarge solution space also. Experiments results show the effectiveness of our approach.