Path: news.mathworks.com!newsfeed-00.mathworks.com!newsfeed2.dallas1.level3.net!news.level3.com!postnews.google.com!news1.google.com!Xl.tags.giganews.com!border1.nntp.dca.giganews.com!nntp.giganews.com!local02.nntp.dca.giganews.com!nntp.supernews.com!news.supernews.com.POSTED!not-for-mail
NNTP-Posting-Date: Mon, 15 Dec 2008 08:56:02 -0600
Newsgroups: comp.soft-sys.matlab
Subject: Re: Which to use (quadprog or fmincon)?
From: "Eric J. Holtman" <ejh@ericholtman.com>
References: <Xns9B72911B0AAD8ejhericholtmamcom@216.168.3.30> <9303149.1229327097884.JavaMail.jakarta@nitrogen.mathforum.org>
Message-ID: <Xns9B755ADC2E3ABejhericholtmamcom@216.168.3.30>
User-Agent: Xnews/5.04.25
Date: Mon, 15 Dec 2008 08:56:02 -0600
Lines: 45
X-Trace: sv3-ipRPP/CPMTLpsyV5eGiozIPFomi3nU8awG+PYi9n+vd0AqzC8URqnWrIrv++IG881Zqn+ongrk2yyOv!EzJ4eTkDlxXjG81jxysaZzaTevNvxZ9P9qTCRGa9orH/u5hCI4R0cnBXk3yxk02Jfvd6ZjV+1iqn!2pOEtdqDsA==
X-Complaints-To: www.supernews.com/docs/abuse.html
X-DMCA-Complaints-To: www.supernews.com/docs/dmca.html
X-Abuse-and-DMCA-Info: Please be sure to forward a copy of ALL headers
X-Abuse-and-DMCA-Info: Otherwise we will be unable to process your complaint properly
X-Postfilter: 1.3.39
Xref: news.mathworks.com comp.soft-sys.matlab:507088


Torsten Hennig <Torsten.Hennig@umsicht.fhg.de> wrote in 
news:9303149.1229327097884.JavaMail.jakarta@nitrogen.mathforum.org:

> 
> 
> The following link might be interesting for you:
> 
> http://plato.asu.edu/sub/nlores.html
> 
> Here a list of codes for the solution of quadratic
> programming problems is given.
> Especially notice the link to the MATLAB code called MINQ.
> 
> Best wishes
> Torsten.
> 


Hi Torsten... thanks for the link, I'll be checking it
over carefully to see if I can find something useful.

MINQ doesn't seem like it will work for me, since it only
allows bounds constraints.  I have inequalities.

I'm still confused by a minor point (I am not, as should
be obvious by now, a mathmetician by trade).

How is it even possible to convert inequalities to equalities,
unless you know that all variables are implicitly >= 0?

It seems to me that the transform of

x1 + x2 <= 40 to    x1 + x2 + __Slack1 = 40 only works
if you know __Slack1 >= 0.  And that seems to me to mean that
either (a) the solver must implicitly force variables to >=0
or (b) it must allow bounds in addition to equalities.

So it seems that if I have inequalities, I can never use
the large-scale quadprog, because I can't force them to 
equalities, because I can't pass bounds.

Is this really true, or am I just missing something?