Premium
A steepest gradient method for optimum structural design
Author(s) -
Templeman Andrew B.
Publication year - 1971
Publication title -
international journal for numerical methods in engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.421
H-Index - 168
eISSN - 1097-0207
pISSN - 0029-5981
DOI - 10.1002/nme.1620030209
Subject(s) - convergence (economics) , gradient method , mathematical optimization , function (biology) , process (computing) , algorithm , mathematics , fibonacci number , computer science , combinatorics , evolutionary biology , economics , biology , economic growth , operating system
This paper describes a search procedure for finding the unconstrained maximum or minimum of a function of many independent variables. If this function represents the volume, cost, stiffness, etc. of a structure and each of the independent variables is associated with a parameter governing the geometry of the structure then designs of optimum geometry may be made in certain cases with the assistance of this search procedure. The procedure consists of steepest gradient calculations used in conjunction with a reverse Fibonacci location process. A ridge‐following technique is included to speed convergence in addition to localized exploration in the region of an optimum. The search procedure has been programmed for computer use and an outline of its structural design applications is presented together with an example of its efficiency in a specific case.