partitioning objective in Metis

Section 5.3 of the Metis manual mentions the two partitioning objectives that can be used with Metis. It seems like it can minimize the total edge cut, or the total communication volume. For some of the graphs that we were trying this on, it leads to some partitions having too many edges that are cut, while other partitions had few edgecuts. I want to "load balance" the edgecut among the different partitions.

It seems like I need the objective function to be
totalv = Max(Nadj[v])
instead of
totalv = Sum(Nadj[v])
And then choose a partition that minimizes the objective function totalv. Is there a way to do this through the executables or the partitioning API?

Thanks in advance.

RE: No you cannot achieve this

No you cannot achieve this objective in Metis, as it will require to change the core refinement routines.