It is now widely recognized that structural optimization using mathematical programming techniques can be employed efficiently only in conjunction with explicit approximate model. In this work an efficient optimization methodology combining finite element-based approximate analysis model, a sequential quadratic programming algorithm incorporating an active set strategy and direct method of design sensitivity analysis is developed. The methodology involves the solution of a sequence of explicit high-quality approximate problems subject to move limits (region of trust) in the design space. A new technique for constructing approximation functions with high quality and adaptive capability to original functions is proposed by using the values of the state variables (displacements and/or stresses) and their derivatives at points obtained in the process of optimization. Other approximation techniques have been presented and comparisons have been made using real-life automobile structures to demonstrate the power and generality of the approximation concepts in structural optimization.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £60 +P&P)