用于结构优化的一种新的两点近似函数

 提出一种基于两点信息的近似函数，它在当前设计点能够保证原函数的函数值、一阶导数值和近似二阶导数值；在另一设计点保证原函数的一阶导数值。用该近似函数形成结构优化原问题的序列近似问题，再用约束变尺度法解近似问题。通过典型算例的计算，结果表明本方法具有较好的收敛性和较高的计算效率，适用于结构优化问题的求解

A NEW TWO POINT APPROXIMATION FUNCTION FOR STRUCTURAL OPTIMIZATION

A new and improved two point constraint approxmiation is presented, which uses the information of the exact function valueits first order derivatives and approximate second order derivatives at the current design point, and also the first order derivatives at the previous point of design iterations. The improved approximate expressions are proposed to create the sequential approximation problems which are solved with the constrained variable metric method. This approximation is efficient and suitable for solving structural optimization problems. And it is proved by some well known examples.