适应任务变化的电磁探测卫星动态调度模型及算法

The Model and Algorithm of Dynamic Scheduling for Electromagnetic Detection Satellite Adapted to Task Change

电磁探测卫星是一种对地观测卫星,其搜集的信息对工业、科研、军事等领域有着重要的意义。目前对于卫星调度问题的研究多是假设调度一旦开始,则参与调度的任务与资源将不再变化。实际上,卫星工作在充满随机扰动的复杂环境中,卫星资源可能出现故障,新任务可能随机到达。如果卫星调度无法适应这些变化,必然导致规划有效性和资源利用率降低。论文针对新任务随机到达的情况,结合电磁探测卫星约束特点,建立了基于条件约束满足问题(CCSP)的多星动态调度模型,提出了基于记忆效应Squeaky-Wheel Optimization(SWO)的电磁探测卫星多星动态调度算法。根据模拟的数据进行实验及分析,表明该方法能有效解决新任务随机到达的电磁探测卫星多星动态调度问题。

Electromagnetic Detection Satellite （EDS） is a type of Earth Observation Satellites. The Information collected by EDS is very important in some application domain, i. e. , industry, science and military. Current research mainly put focus on the satellites scheduling in a static fashion, namely, making a hypothesis that when the scheduling start, the resource and task is immutable. In actual, the satellites are working in a complex environment,faced with different uncertainties and perturbations such as unanticipated changes in satellite resources, or arriving of new tasks. It is then necessary to schedule the Earth Observing Satellites in a dynamic way. Considering the specific constrains of EDS, a Conditional CSP Multi-Satellites scheduling model was established in the condition of new tasks arriving randomly and an algorithm based on Squeaky-Wheel Optimization （SWO） was proposed. Finally, some experiments were implemented to valid correctness and practicability of our algorithm.