航空学报 > 2021, Vol. 42 Issue (4): 525060-525060   doi: 10.7527/S1000-6893.2021.25060

考虑高耗时约束的追峰采样智能探索方法

龙腾1,2, 毛能峰1, 史人赫3, 武宇飞1, 沈敦亮4   

  1. 1. 北京理工大学 宇航学院, 北京 100081;
    2. 北京理工大学 飞行器动力学与控制教育部重点实验室, 北京 100081;
    3. 清华大学 航天航空学院, 北京 100084;
    4. 北京宇航系统工程研究所, 北京 100076
  • 收稿日期:2020-12-07 修回日期:2020-12-30 发布日期:2021-02-02
  • 通讯作者: 龙腾 E-mail:tenglong@bit.edu.cn
  • 基金资助:
    国家自然科学基金(51675047,52005288);航空科学基金(2019ZC072003)

Mode pursuing sampling intelligent exploring method considering expensive constraints

LONG Teng1,2, MAO Nengfeng1, SHI Renhe3, WU Yufei1, SHEN Dunliang4   

  1. 1. School of Aerospace Engineering, Beijing Institute of Technology, Beijing 100081, China;
    2. Key Laboratory of Dynamics and Control of Flight Vehicle of Ministry of Education, Beijing Institute of Technology, Beijing 100081, China;
    3. School of Aerospace Engineering, Tsinghua University, Beijing 100084, China;
    4. Beijing Institute of Astronautical Systems Engineering, Beijing 100076, China
  • Received:2020-12-07 Revised:2020-12-30 Published:2021-02-02
  • Supported by:
    National Natural Science Foundation of China (51675047, 52005288); Aeronautical Science Foundation of China (2019ZC072003)

摘要: 针对现代飞行器设计等工程优化问题中面临的约束高耗时难题,在标准追峰采样(MPS)方法的基础上,提出了一种基于过滤器的MPS-DCP设计空间智能探索方法(FMPS-DCP),训练径向基函数网络预示高耗时目标函数与约束条件响应,利用KS方程聚合高耗时约束并根据过滤器思想筛选优质简单样本点,定制了一套新增样本点选择策略引导优化过程快速向全局可行最优解收敛,从而提高了求解高耗时约束优化问题的效率。采用一组标准约束测试算例验证FMPS-DCP方法的性能,并与CiMPS、Extended ConstrLMSRBF、ARSM-ISES和KRG-CDE智能探索方法进行对比。结果表明,FMPS-DCP在优化效率与鲁棒性方面具有显著的性能优势。最后,通过全电推进卫星平台多学科设计优化案例,验证了FMPS-DCP的工程实用性。

关键词: 追峰采样, 智能探索方法, 高耗时约束, 近似优化, KS方程, 过滤器

Abstract: The engineering optimization practices such as modern flight vehicle design often encounter expensive constraints. Based on the standard Mode Pursuing Sampling (MPS) method, a Filter-based Mode Pursuing Sampling intelligent exploring method using Discriminative Coordinate Perturbation (FMPS-DCP) is proposed in this work for constrained optimization problems. In this work, the radial based function network is trained for predicting the values of expansive objective function and constraint functions, and KS function is used to aggregate constraints. Then a filter is constructed for deciding whether to accept sampling points, and a sample point selection strategy is designed to lead the algorithm converge to global feasible optimal value rapidly. FMPS-DCP is tested on a number of standard numerical benchmark problems and compared with CiMPS, Extended ConstrLMSRBF, ARSM-ISES and KRG-CDE. The optimization results indicate that the optimization efficiency of FMPS-DCP is higher than others with lower standard deviation for multiple runs. Finally, the practicality of FMPS-DCP is demonstrated by an all-electric propulsion satellite platform multidisciplinary design optimization problem.

Key words: mode pursuing sampling, intelligent exploring method, expensive constraints, approximate optimization, KS function, filter

中图分类号: