地质灾害应急监测的卫星调度算法研究

This project will focus on the research of theoretical model, effective algorithm of multi-satellites collaborative scheduling and the evaluation method of the optimization results for emergency monitoring of geological hazards, which is a NPC problem. There are many theoretical challenges, such as multi-constraint, multi-objective and high time-effectiveness, for solving the problem. Our purpose is to exploit some new methods and techniques by applying planning and scheduling theory, multi-objective optimization theory and the manifold learning theory, which are effective for emergency monitoring of geological hazards. To meet these specific requirements for satellite resources asked by emergency monitoring for the different geological hazards and the disaster-causing factors, a multi-objective scheduling model for emergency monitoring of geological hazards will be established. Then, the guidance and aim of searching direction for multi-objective optimization in the process of searching optimal result are researched, and an effective scheduling algorithm which meets the requirements of real-time will be proposed based on the theory of manifold learning. Meanwhile, in order to solve the problem that the decision makers fail to make an effective decision by using the optimal results of multi-objective optimization algorithm, the hierarchical evaluation index system for the optimal results of multi-objective optimization will be set up. An effective method for assessing each individual in the optimal results of multi-objective optimization algorithm will also be put forward. The research of this project will significantly improve the effectiveness of the data acquisition in geological hazards and provide a solid support for gaining spatial data in disaster relief and disaster alleviation.

本项目着重研究地质灾害应急监测时多星协同调度的理论模型、快速算法及优化结果的评价方法。目的在于运用规划调度理论、多目标优化理论、流形学习理论开辟一些适用于地质灾害应急监测的新方法与新技术。地质灾害应急监测的多星协同调度是一类NPC问题，其求解面临着多约束、多目标和强时效性等理论难题。针对不同类型地质灾害及孕灾因子的监测对卫星资源的特定需求，本项目拟建立一种面向地质灾害应急监测的多目标调度模型，并基于流形学习理论，研究多目标优化算法寻优过程中搜索方向的引导性与目的性，提出一种满足实时性要求的快速调度算法。同时，针对多目标算法的优化结果往往不利于应急时决策者及时做出有效决策的问题，本项目拟建立多目标算法优化结果的分层评价指标体系，并提出一种有效的多目标算法优化结果的评价方法。本项目的研究必将显著改善地质灾害发生时灾区数据获取的时效性，为救灾减灾提供坚实的空间数据支持。

针对地质灾害的突发性与应急救灾的时效性，如何有效整合卫星资源，进行统一管理和规划调度，实现卫星资源的协同工作，充分发挥对地观测尤其是应急监测的效能，从而实时获取灾区的遥感影像数据，以在重大地质灾害等突发情况下对灾害区域进行有效的观测，形成高效的应急救助体系，是一个亟待解决的重要问题。.地质灾害应急监测中的成像卫星调度问题具有强时效性要求，有多个相互冲突的目标要同时进行优化，是一类具有多峰多谷的多变量、多约束、多目标的复杂计算问题。项目针对地质灾害应急监测卫星调度模型中的多目标及高维问题特性，通过关联操作融合了基于指标的多目标演化算法和基于参考点的多目标演化算法，应用基于参考点的密度选择机制，提出了一个新的基于增强θ支配和密度选择的演化算法，改进了多目标优化算法的收敛性与多样性。.地质灾害应急监测多星协同调度问题本质上是一种有限资源调度问题，面临着各种复杂的约束，从而导致问题的复杂性。项目将多星联合任务协同规划问题抽象为一类常规的带复杂时间窗口约束的多资源联合调度规划问题，构建了多种覆盖要求下的混合整数线性规划模型，面向地质灾害应急监测环境下的动态任务特性，提出了自学习策略控制下的星上资源自主任务分配方法、面向大规模任务请求的多星联合调度智能优化算法、动态应用需求下的资源感知和星上任务快速自主规划方法。.重大地质灾害发生时，应急救灾指挥部通常都需要根据应急调度的结果及时做出应急决策，这就要求对调度算法的结果进行科学评判，便于应急决策方案的制定与选取。项目深入地研究了各种调度方案下的对地观测覆盖计算方法，提出了卫星对地覆盖计算的形式化数学描述方法，提出了经度条带法、纬度条带法等多种对地覆盖的快速计算方法，提出了多星联合对任意类型地面目标区域的精确覆盖计算方法。.项目的研究成果改善了地质灾害发生时灾区数据获取的时效性，为救灾减灾提供了坚实的空间数据支持。