[Coin-discuss] Benchmark Instances & code for format conversions

[Michael Hennebry]

On Thu, 27 Apr 2006, Tobias Achterberg wrote:

> Dear Anureet,

> > To summarize, availability of benchmark instances in "appropriate
> > format" is an equally important aspect of computational OR as is the
> > availability of the code. In my opinion, maintaining benchmark instances
> > and "format conversion codes" along with the implementation code would
> > greatly enhance the scope and reach of COIN-OR.
>
> I totally agree. But please keep in mind that this is a non-trivial
> task. Just creating a website where everyone can upload instances is

More important than having an appropriate format
is having a format that is well documented.

MPS is not necessarily an appropriate format.
It won't work for nonlinear problems with
a nonlinearity other than integrality.
It's not good for TSPs.
If one is testing a CTP solver,
having it read a format that could represent
any MILP would not seem to be a good idea.

> by far not enough. You will have to actively collect instances, make
> sure that you give credit to the original source, if file format
> conversion has to be applied you have to indicate this and give the
> instance a different name, ....

Another possibility is to insist on really good format documentation.
If the format defines a MILP, one could insist that
the documentation include source for library functions that can
read a file into a data structure and
given the corresponding variable name or row name,
return a representation of the exact value of any
of the coefficients or bounds defining the MILP.
The format of the exact representation should be defined by COIN-OR.

-- 
Mike   hennebry at web.cs.ndsu.NoDak.edu
"Demons after money?
Whatever happened to the still-beating heart of a virgin?
No one has any standards any more."                    --  Rupert Giles