Re: Solving linear system of equalities AND disequalities

pmatos wrote:
> It seems to me that the method you proposed is very inefficient. If you
> have n disequalities that will generate an enormous amount of branches
> (2^n).

I should have made it more clear that you only split on each case when
necessary. There are a number of approaches based on SAT solvers that
would eliminate the need to split on each of the 2^n cases.
Unfortunately, I am not aware of a tool that does this already, so I
think it is still is a research problem.

> I'm thinkin' about lazy evaluation of disequalities. Probably ignoring
> disequalities and requesting for a solution, whenever a solution if
> found you check to see if the disequalities allow the solution, if they
> do output, else find other solution. Of course you'd have to check for
> cases where you have x+y = 1 in equalities and x+y != 1 in
> disequalities! :D

What happens if there are an infinite number of solutions?
.