火星探测转移轨道的鲁棒优化设计  

Robust optimization design for Mars exploration orbit

作　　者：余艳[1] 戴光明[1] 王茂才[1] 宋志明[1] 

机构地区：[1]中国地质大学(武汉)计算机学院,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2015年第8期88-91,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61103144;61472375);中国博士后科学基金资助项目(2012T50681)

摘　　要：以设计变量存在不确定的火星探测轨道设计模型为研究对象,采用最小最大鲁棒优化方法获得该模型的鲁棒优化解.在传统的最小最大方法的基础上,考虑了决策变量的扰动,提出了一种嵌套的差分演化算法.在该算法中,内层差分演化算法计算不确定域的最差目标函数值,外层差分演化算法获得全局的鲁棒优化解.通过试验对该方法的有效性进行了验证,结果表明:该方法适合于目标函数没有解析表达式、高度非线性的问题,实际工程问题中的不确定性不可忽略.A robust model of Mars exploration transfer trajectory was built from its classical configu‐ration by assuming the imprecisions in the design space .Be aware of the disadvantages of the classical robust optimization of min‐max method ,a novel approach named a nested differential evolution algo‐rithm was applied to solve the robust optimization problem .In the algorithm ,the inner differential e‐volution algorithm was used to find the local worst‐case value in the uncertainty sets and the outer one was applied to obtain the global optimum in the decision space .The behavior of the robust optimum was examined and the efficiency of the algorithm was verified .The results of the numerical calculation show that this approach is applicable for problem which can not be expressed by an analytical closed form or is high nonlinear ,and the imprecisions in the real engineering problems should not be ignored .

关 键 词：火星探测 最小最大方法 鲁棒优化 差分演化 优化设计 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]