# Partitions with bounds on the size of parts

Hello,

I am trying to solve a graph partitioning problem, where I want to find a partition with a minimum edge cut such that the size of each part is not bigger than nmax.

If my understanding is correct, to use METIS' graph partitioning routines, the size of each part should be given as an input. Is there a way to directly use metis to solve this more

general problem? Currently, I am calling metis inside a loop where I solve the partitioning problem with some different set of part sizes ... I wonder if there is a better way to do this!

Many thanks!

Aida

Submitted by aida on Wed, 2012-02-29 17:09

## RE: My Partitioning Problem

Dear Aida

I am trying to solve a graph partitioning problem, where I want to find a partition with a minimum edge cut such that the size of each part is not bigger than nmax.(which was your question)

Also I assigned the resources to each node in order to the number of total resources of each partition not bigger than rmax?

Do you know it is possible to implement these two constraints with METIS or such a like software or in the other way :)?

Thank in advance for your help and considerations

Regards