基于TVNS的敏捷对地观测卫星自主动作规划算法

Autonomous Action Planning of Agile Earth Observing Satellite Algorithm Based on TVNS

针对敏捷对地观测卫星自主动作规划问题,建立了基于时间线约束网络的问题模型。通过对模型的分析,归纳出9种卫星动作序列,确定了各种动作序列之间的衔接关系,并在此基础上设计了动作时间的计算和调整方法,提出了基于三阶段变邻域搜索(Three-stage Variable Neighborhood Search, TVNS)的敏捷对地观测卫星自主动作规划算法。算法以任务调度结果作为初始解,分3个阶段,使用2种邻域结构进行变邻域搜索,实现了任务到动作的转换。通过实验证明了基于TVNS的敏捷对地观测卫星自主动作规划算法的有效性。

To address the autonomous action planning of agile earth observing satellites, a model based on constraint network on timelines is constructed. Through the analysis of the model, 9 action sequences are summed up, and the link relation between different action sequences are confirmed. Then, the calculation and adjustment method of action time is designed;and an autonomous acting planning of agile earth observing satellite algorithm based on Three-stage Variable Neighborhood Search(TVNS) is presented. The algorithm takes the task scheduling result as initial solution, and uses two neighborhood structures to conduct variable neighborhood search by three-stage, thus achieve the transformation from mission to action. Experimental results show the effectiveness of autonomous acting planning of agile earth observing satellite algorithm based on TVNS.