›› 2012, Vol. 32 ›› Issue (1): 77-83.doi: 10.3780/j.issn.1000-758X.2012.01.012

• 技术交流 • 上一篇    

航天器模块划分数值优化方法

刘永健, 刘育强, 石军, 谭春林   

  1. (中国空间技术研究院,北京100094)
  • 收稿日期:2011-05-19 修回日期:2012-02-25 出版日期:2012-02-25 发布日期:2012-02-25
  • 作者简介:刘永健 1974年生,2003年获清华大学固体力学专业博士学位。主要从事航天器系统研发工作。

Numerical Optimization for Spacecraft Module Identification

LIU  Yong-Jian, LIU  Yu-Qiang, SHI  Jun, TAN  Chun-Lin   

  1. (ChinaAcademyofSpaceTechnology,Beijing100094)
  • Received:2011-05-19 Revised:2012-02-25 Online:2012-02-25 Published:2012-02-25

摘要: 对航天器模块划分主要依靠经验和定性方法的问题,构建了一种通用的模块划分数值优化方法。利用功能分解方法、相关性评价准则、设计矩阵与设计结构矩阵(DesignStructureMatrix,DSM)同步演化的机制获得了基于功能元的数值DSM;基于DSM构建了遵循公理化设计原理的通用优化模型;利用遗传模拟退火算法获得了可对模块划分方式及数目进行同步优化的模块划分方法;给出了不同于优化目标的评价方法。仿真结果表明,该方法可将航天器划分为一系列内聚度高、耦合度低的模块。

关键词: 设计结构矩阵, 遗传算法, 模拟退火算法, 模块划分, 航天器

Abstract: To overcome the problem that spacecraft module identification mainly relies on experience and qualitative methods,a general numerical optimization for module identification method was constructed. Function units based numerical design structure matrix (DSM) was obtained by using functional decomposition method, relativity evaluation criteria and coevolution mechanism between design matrix and DSM. A general DSM based optimization model was constructed which conformed to axiomatic design axioms. A genetic-simulated annealing algorithm was used to construct a module identification method which could co-optimize the mode and the number of module identification. Finally, an evaluation method for different optimization targets was put forward. Simulation results show that the proposed method can divide a spacecraft into several modules with high cohesion degree and low coupling degree.

Key words: Design structure matrix, Genetic algorithm, Simulated annealing algorithm, Module identification, Spacecraft