Re: [Help-glpk] The "good enough" set covering problem

[Yaron Kretchmer]

Got it.
What influences the runtime - the number of sets, or the size of the domain? 

The reason I'm asking is that it's pretty easy to break the domain up into smaller domains, find good coverages for each of the smaller domains, then take a union of all the resulting minimum sets. Optimality is not a requirement here, just a "good enough" solution.

Regards
Kretch

On Mon, Feb 9, 2009 at 11:47 AM, Andrew Makhorin <address@hidden> wrote:

> Regarding the problem size- are we talking about memory limitations? 

> (I #39;ll be using a 64bit machine with 32G RAM)

No. The glpk mip solver may take too long time to solve your problem
to optimality.