The local approximation method for structural optimization

Yi Nie, Yan Wang, Wei Sun, Yan He, Jing Hao, Tao Huang

Research output: Chapter in Book/Conference proceeding with ISSN or ISBNConference contribution with ISSN or ISBNpeer-review


The local approximation method exhibits many advantage features and it is popular to a broad class of structural optimization problems. In this paper, both the mathematical modeling and case study of the local approximation method were studied. The theoretical analysis indicates that the primary optimization problem can be replaced with a sequence of explicit approximate problems by using the local approximation method. The explicit subproblems are convex and separable, which can be solved efficiently by using a dual method approach. The topology optimization of a guide rail design is then solved to testify the proposed method, which has been coded by Altair OptiStruct. The optimized design of a widely used guide rail with an "I" shape cross section is obtained and compared with the original design. The numerical results have shown that the local approximation method can effectively solve the structure optimization problems, especially the ones with hundreds of design variables or constraints.

Original languageEnglish
Title of host publicationMaterials Engineering and Automatic Control III
PublisherTrans Tech Publications Ltd
Number of pages5
ISBN (Print)9783038351405
Publication statusPublished - 1 Jan 2014
Event3rd International Conference on Materials Engineering and Automatic Control, ICMEAC 2014 - Tianjin, China
Duration: 17 May 201418 May 2014

Publication series

NameApplied Mechanics and Materials
ISSN (Print)1660-9336
ISSN (Electronic)1662-7482


Conference3rd International Conference on Materials Engineering and Automatic Control, ICMEAC 2014


  • Local approximation method
  • Mathematical modeling
  • Structural optimization
  • Topology optimization


Dive into the research topics of 'The local approximation method for structural optimization'. Together they form a unique fingerprint.

Cite this