Re: [Help-glpk] Podhradsky (2010) review of optimization modeling languages

[Nigel Galloway]

Robbie,

Are you telling me that an electrical engineer, to be, should or should
not base his thesis on a comparison of modelling languages?

Do you agree with the papers conclusion that it is, and am I supposed to
read it as:
 'A thorough comparison of available modelling languages for
 Mixed-Integer Linear Programming'?

He seems to have looked at
http://www.orms-today.org/surveys/LP/LP-survey.html (we can read excel
spreadsheets can we not?) and
http://www.orms-today.org/surveys/LP/lp9.html. From which he concludes
that mathprog has fewer features. Do you think this is sufficient?

Can the model he proposes be written in Mathprog? Can it be solved by
GLPK?

Can you explain what the conclusion 'As the best performance was made by
Zimpl, a set-oriented modelling language, it is reasonable to expect
that other set-oriented languages give a similar performance.
Nonetheless, this assumption was not proved.' means? - The good news is
Mathprog is set oriented, but why should a  matrix oriented language be
less efficient? 

Is there not a lot of ambiguity for a comprehensive conclusion in his
final conclusion:
 'Zimpl generates MPS/LP files whereas the other two languages call
 solvers using an embedded API. Loading
 large models stored in MPS/LP files by appropriate solvers might be
 time-consuming and not as fast as
 using a solver API embedded into the modelling language. Another
 interesting question is the speed of
 OptimJ extraction time for MPS or LP files (for Zimpl the performance
 for both file formats is 
 identical). It is also questionable whether there is a possible
 difference in solver computation time if
 different MILP problem formats are used (e.g. an MPS or LP file
 format).'

Is loading large MPS files into GLPK time-consuming? Does GLPK’s
computation time depends upon which MILP problem format is used?

Perhaps languages for LP/MIP require a Chrestomathy so that we can do
this research properly.

-- 
  Nigel Galloway
  address@hidden


On Fri, Jan 20, 2012, at 11:09 PM, Robbie Morrison wrote:
> 
> Hello all
> 
> Those interested in a comparison between modeling
> languages and translation performance may be interested
> in Podhradsky (2010).  GLPK MathProg is not considered
> by AMPL is.
> 
>     Podhradsky, Michal.  2010.  Modelling languages for
>         optimization -- Bachelor Thesis.  Czech
>         Technical University in Prague, Czech Republic.
>         (the 'y' in 'Podhradsky' is unicode U+00FD)
> 
>     
> http://support.dce.felk.cvut.cz/mediawiki/images/0/0b/Bp_2010_podhradsky_michal.pdf
> 
>     abstract (first sentence): The aim of this work is to
>     thoroughly compare various currently available
>     modelling languages for mixed-integer linear
>     programming (MILP), both commercial and open- source,
>     and eventually choose and recommend the one which is
>     the most suitable for the task of modelling and optimal
>     scheduling of cogeneration systems.
> 
> best wishes, Robbie
> ---
> Robbie Morrison
> PhD student -- policy-oriented energy system simulation
> Institute for Energy Engineering (IET)
> Technical University of Berlin (TU-Berlin), Germany
> University email (redirected) : address@hidden
> Webmail (preferred)           : address@hidden
> [from Webmail client]
> 
> 
> 
> _______________________________________________
> Help-glpk mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/help-glpk
> 

-- 
http://www.fastmail.fm - The way an email service should be