论文部分内容阅读
针对采用天气预报的滞后云层进行卫星调度影响观测图像质量和观测收益的问题,提出一种获取实时云层的数学模型,并基于此构建考虑实时变换云层的敏捷观测卫星(AEOS)调度模型。由于贪婪搜索算法(GSA)具有局部优化的特性,能够充分考虑卫星观测的云层和有限存储资源等约束,研究了GSA在该卫星调度问题中的应用。首先,GSA优先考虑观测任务的云层遮挡,并根据云层遮挡大小,计算待观测任务的图像质量,将之排序选择待观测的任务;其次,结合任务的大小、截止时间和卫星的存储资源约束,选择能够给观测收益带来最大化的任务;最后,进行观测和任务传送。仿真实验表明,在任务数为100的情况下,采用GSA进行卫星调度的任务收益比常用于卫星调度的动态规划算法(DPA)所获得任务收益提高了14.82%,比局部搜索算法(LSA)所获得任务收益提高了10.32%,并且同等条件下,采用GSA得到的观测图像的质量比其他两种方法得到的图像质量更高。实验结果表明,GSA在实际卫星调度中,能够有效地提高图像观测质量和任务观测收益。
Aiming at the problem that the satellite dispatching affected the observed image quality and the observed revenue by the satellite weather forecasting, a mathematical model for acquiring real-time clouds is proposed and an AEOS scheduling model considering real-time transform clouds is constructed. Due to the local optimization of greedy search algorithm (GSA), the GSA can be applied to the satellite scheduling problem by fully considering the constraints of cloud observation and limited storage resources. Firstly, the GSA gives priority to the cloud cover of the observation task and calculates the image quality of the task to be observed according to the size of the cloud cover, and then selects the tasks to be observed. Secondly, combining the task size, deadline and satellite resource constraints, Choose the tasks that maximize the observed benefits; and finally, observe and deliver the mission. The simulation results show that the mission gain of satellite scheduling using GSA is 14.82% higher than that of dynamic scheduling algorithm (DPA), which is commonly used in satellite scheduling, when the number of tasks is 100, which is higher than that of LSA The gain of the task was increased by 10.32%, and under the same conditions, the quality of the observed image obtained by GSA was higher than that of the other two methods. The experimental results show that GSA can effectively improve the quality of image observation and mission observation in the actual satellite scheduling.