From: Niko Vuokko on
Is there a way to use a sparse matrix as A (rule coefficient matrix) in bintprog? It complains that the inputs should be double (they are, A is just sparse). It is utterly impossible to use a dense matrix as A, because I have less than a terabyte of memory in my machine :)
From: John D'Errico on
"Niko Vuokko" <ntvuok(a)gmail.com> wrote in message <g9t5cu$kgv$1(a)fred.mathworks.com>...
> Is there a way to use a sparse matrix as A (rule coefficient matrix) in bintprog? It complains that the inputs should be double (they are, A is just sparse). It is utterly impossible to use a dense matrix as A, because I have less than a terabyte of memory in my machine :)

My guess is that even if you were able to
provide a sparse matrix that large, then you
would not see bintprog return a solution to
that problem in your lifetime.

John
From: Niko Vuokko on
"John D'Errico" <woodchips(a)rochester.rr.com> wrote in message <g9tgkt$4t1$1(a)fred.mathworks.com>...
>
> My guess is that even if you were able to
> provide a sparse matrix that large, then you
> would not see bintprog return a solution to
> that problem in your lifetime.
>
> John

I know that and I would use the MaxTime option. The matrix only has about a few million non-zeros, so it's very sparse. There are also easier ways to get results, but this would be an important comparison and for comparison I need actual results, "not realistic" isn't enough.