[Coin-discuss] Suggestions on how to go about tuning?

[Michael Hennebry]

On Sat, 3 Jun 2006, John J Forrest wrote:

> The first thing to do is solve the problem however slowly and look at log
> and solution.  Using defaults this took 1148 seconds and 150K nodes.
> Looking at the log file it looked as if strong branching was not buying
> very much.  What was more interesting was that all variables took integer
> values although only half were declared integer.  Assuming this would
> always be true I tried setting them all integer.  This time strong
> branching was kicking in and it took 5K nodes and 413 seconds.  I also saw
> that cuts and heuristics were not doing much good so switched them off.

It occurs to me that sometimes it might be important to be able
to distinguish between constraints that define the problem
and constraints that can be inferred or can be accepted.

In the case at hand, at least one solver runs faster and gets
the right answer with all variables specified as integer.

'Twouldn't be a surprise if some solver ran faster with
only the original integer variables declared integer.

A solver that knew which kind it was could accept or reject
constraints that didn't define the problem accordingly.
Better yet, it might use a strategy that enforces or relaxes
a non-problem-defining constraint whichever is more efficient.

In the case at hand, a solver might employ a rounding heuristic
that did not enforce integrality on all the variables,
but it might nevertheless use Gomory fractional cuts,
which require integer variables.

-- 
Mike   hennebry at web.cs.ndsu.NoDak.edu
"it stands to reason that they weren't always called the ancients."
                                                      --  Daniel Jackson