Hi Hans,<br><br><div class="gmail_quote">On Thu, Jun 7, 2012 at 9:47 PM, Hans Pirnay <span dir="ltr"><<a href="mailto:hans.pirnay@rwth-aachen.de" target="_blank">hans.pirnay@rwth-aachen.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi Novalio,<br>
<div class="im"><br>
> I would like to know if windows binary of sIPOT is available for download in<br>
> internet or not. I have zero experience in compiling C/C++ code on windows.<br>
> I just want to have a try and see how it can be used in my work. What I need<br>
> this time are:<br>
<br>
</div>there are no sIPOPT binaries available for download. However, sIPOPT<br>
is a library (and an AMPL executable) that link against the Ipopt<br>
library. The biggest difficulty in compiling Ipopt on Windows systems<br>
are the Fortran libraries. If you have Ipopt working, this problem has<br>
been solved already, and compiling sIPOPT should not be a big deal.<br>
<div class="im"><br>
><br>
> 1. sensitivity factor of objective function with respect to a variable<br>
> change.<br>
> 2. sensitivity factor of constraint function with respect to a variable<br>
> change.<br>
<br>
</div>What exactly do you mean by sensitivity factor?<br>
<br></blockquote><div><br>I mean the first derivative only, the Jacobian. If I undestand it correctly, sIPOPT can be used for the first one. For the second one, I still do not have any idea.<br> </div><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
sIPOPT treats problems of the form<br>
<br>
min_x f(x,p)<br>
s.t. c(x,p) = 0<br>
x>=0<br>
<br>
where x are decision variables and p are parameters, which are fixed<br>
for the run of the optimization problem.<br>
sIpopt then aides in computing several things, among them<br>
<br>
- the sensitivity of the decision variables w.r.t. the parameters dx/dp<br>
- given a nominal solution x(p_0), compute x(p) for small |p-p_0|<br>
<br>
There is a paper describing the capabilities of sIPOPT at<br>
<a href="http://www.optimization-online.org/DB_HTML/2011/04/3008.html" target="_blank">http://www.optimization-online.org/DB_HTML/2011/04/3008.html</a><br>
<br></blockquote><div><br>I think it is important for me to read your paper soon. Only after that I can decide to use sIPOPT or not.<br> <br></div><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
Also note that sIpopt has recently been re-implemented from scratch.<br>
The code can be found at<br>
<a href="https://github.com/athrpf/sipopt2" target="_blank">https://github.com/athrpf/sipopt2</a><br>
this new code introduces parameters explicitly into Ipopt. If you are<br>
seriously considering using sIPOPT, you should look into the new<br>
version. Among many other technical advantages, its AMPL interface is<br>
much easier to use.<br>
<br>
Hope this helps<br>
<span class="HOEnZb"><font color="#888888">Hans Pirnay<br>
</font></span></blockquote></div><br>Best regards.<br><br>Novalio Daratha<br>