[eclipse-clp-users] Teorethical Question

Hi,

 

I am dealing with a number of linear constraints (linear inequalities) and I
am curious if exists some theoretical results regarding the minimum number
of constraints which can be violated having as an input the inequalities and
the domains of the variables. 

>From what I read on the Internet I found articles like this:
http://www.springerlink.com/content/fr0c10gxnbm0ql93/

Do you know similar work for the particular case when the constraints are a
linear combinations of variables?

 

Thank you,

Bogdan.