[Fwd: Tricks with MathProg to approximate non-linear functions?]

-------- Forwarded Message --------
From: Dolan Antenucci <[email protected]>
To: [email protected]
Subject: Tricks with MathProg to approximate non-linear functions?
Date: Fri, 05 May 2017 16:37:43 +0000

I'm attempting to use GLPK to solve a problem with a non-linear
objective function. Specifically, I want to use either a Pearson
correlation coefficient
(https://en.wikipedia.org/wiki/Pearson_correlation_coefficient) or
something similar to the F1 score metric
(https://en.wikipedia.org/wiki/F1_score). 



I know that GLPK is restricted to *linear* programming, but I'm
wondering if there is a trick to representing either of these objectives
as linear functions.  


I got some hope when I came across a guide for "MIP linearizations and
formulations" from FICO
(http://www.fico.com/en/node/8140?file=5125), which talks about
approximating non-linear functions with a piecewise linear function, but
since it is in regards to their Xpress Optimization Suite, I wasn't sure
how this applied to my case or with GLPK.


Are there any known tricks with GLPK for what I'm trying to do, or am I
best off just choosing a linear objective function?


Best Regards,
Dolan Antenucci