We study the problem of finding the minimum bisection of a graph into
two parts of prescribed sizes. We formulate two lower bounds on the pr
oblem by relaxing node- and edge-incidence vectors of cuts. We prove t
hat both relaxations provide the same bound. The main fact we prove is
that the duality between the relaxed edge- and node- vectors preserve
s very natural cardinality constraints on cuts. We present an analogou
s result also for the max-cut problem, and show a relation between the
edge relaxation and some other optimality criteria studied before. Fi
nally, we briefly mention possible applications for a practical comput
ational approach.