[Coin-lpsolver] how large scale problems clp can solve?

[Jinwei Gu]

Dear All,

I am new to CLP. I am wondering what is the limit of the scale of the
problems that CLP can solve.

My research problem has about 1 million unknowns (i.e. columns), and 4
million constraints (i.e. rows). The matrix is very sparse. Each of the rows
only has about 100~400 non-zero values. I am wondering whether CLP can be
used to solve this problem?

I ran several test problems and CLP works much faster than glpk which I used
before. Actually CLP is the only open-source library I found that might be
able to solve that large scale problems. Is it just limited by my machine
memory? Is there a difference in terms of the scale limit, if I use simplex
method (primal/dual) or the barrier method to solve?

I searched through the mailing list archive but could not find suitable
answer. Thanks a lot for your help.


-- 
Best wishes,
Jinwei Gu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/clp/attachments/20080127/37841771/attachment.html>