We will scale a n question, obtained by the partition of a one size of $ \ FRAC {n} {b} $ sub-problems, each recursive additional computing brings to f (n) , then we get the following relationship formula:
$T(n)=aT(\frac{n}{b})+f(n)$,
In addition, we define a $ c_ {Crit} $ , so it is calculated:
1. When $ f (n) = O (n ^ c), and c <c_ {crit} $:
$T(n)=\Theta(n^{c_{crit}})$
2. When $ f (n) = O (n ^ c), and c = c_ {crit} $:
$T(n)=\Theta(n^{c_{crit}}log{n})$
3. When $ f (n)> O (n ^ c), and c = c_ {crit} $:
$T(n)=\Theta(f(n))$
Reference links: