Assume an arbitrary r
T(n) <= cn + T(n/r) + T (3n/4)
show T(n) <= Dcn
for some constant D
by reworking the induction proof, use the expression to argue that:
T(n) <= Dcn
does not hold for r=3
.
Assume an arbitrary r
T(n) <= cn + T(n/r) + T (3n/4)
show T(n) <= Dcn
for some constant D
by reworking the induction proof, use the expression to argue that:
T(n) <= Dcn
does not hold for r=3
.
Have a look at the Akra-Bazzi theorem. This is a generalization of the master theorem that does not require subproblems of equal size.