ct2147 wrote:
Hi, there,

I am using Excel VBA programming to do the optimization, and I already
finished my first step using Excel solver, but the problem is that Excel
solver has to renew the worksheet to find the solution, and to speed up for
optimization purpose, I wonder if anyone could offer me some help in solving
the problem to proceed in the 2nd step: to replace Excel solver with some
free solver codes to speed up the whole process .

Specifically, the problem is depicted as finding a root solver:

we have variable (x1,x2,x3...,xn), and function f(x1,x2...,xn) =
(y1,y2...,yn), we want to call another function to find a set of (x1,...xn)
to make (1y1,y2,...yn) equal to 0. (n less than 100)

Alternatively, this problem can be converted to another problem: miminizing
the squar sum of (y1,y2...,yn) also leads to the same results. So we need a
minimization function. The key point is we can only provide the function, not
the derivatives of the function.

The ideal solution is to find some solver codes from free source in C or
Fortran (any help), and then I can compile them into a *.dll library file and
use them in the Excel VBA.

The NEOS Guide Optimization Tree
http://www-fp.mcs.anl.gov/otc/Guide/OptWeb/
has links to many codes in C, C++, and Fortran. A good newsgroup for
optimization questions, including suggestions of which code to use, is
sci.math.num-analysis.