[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Help-glpk] We live in a world of hidden inequalities
From: |
Andrew Makhorin |
Subject: |
Re: [Help-glpk] We live in a world of hidden inequalities |
Date: |
Thu, 05 May 2011 17:25:05 +0400 |
> Don't worry, I think this has nothing to do with global economics or
> worldwide peace.
>
> I have a model which, when run, glpk advises me that it has discovered
> 162 hidden inequalities. I assume in the model and not the world,
> because 162 would seem small for the world. Also the numbr varies with
> the data I supply to the model and its difficult to see how that would
> affect the world.
>
> What are they? Both in a general sense, what are hidden inequalities?
> and in the particular case of my model, how do I investigate them?
>
> Are they a good or a bad thing? The model works quickly and produces
> correct answers. Should every model have some? How have I induced them?
>
> LPs for a long time have had only two objectives, min and max, and they
> are really the same. Perhaps it is time for a new objective to find the
> data with the minimum hidden inequalities, which is presumably the
> fairest solution.
>
Cover inequality is an inequality in the form
x1 + x2 + ... + xn >= 1
where x1, x2, ..., xn are binary variables. It means that at least one
xj should be 1 in any integer feasible solution. Hidden cover inequality
is an inequality which doesn't have the form above, but is equivalent to
some cover inequality. For example,
2x1 + 3x2 + 2x3 >= 2
is hidden cover inequality, because it is equivalent to and therefore
can be replaced by the following cover inequality:
x1 + x2 + x3 >= 1
Such replacements are used internally by the mip preprocessor, and you
may just ignore all messages concerning details of the preprocessing.