Show that as a consequence of there being

a finite number of possible assignments for the set of discrete indicator variable rnk$,$ and that

for each such assignment there is unique optimum for the $\{$k$\},$ the K$-$means algorithm must

converge after a finite number of iterations.