When is $n^β > \log^α n$?

39 Views Asked by At

I've been doing asymptotic time complexity analysis, and someone made me notice that although we say that:

$$\log^α n = o(n^β)$$

For all $β>0$, given a smallish $β$ and a moderate $α$, it actually takes a very large $n$ for:

$$n^β > \log^α n $$

To hold. For example, the inequality:

$$n^{1\over 6} > \log^2 n$$

Holds only from around $n > 8.79942×10^{19}$ (excluding a short interval for some $n < 3$).

How do I solve this kind of inequality?

I understand the solution has to do with the Lambert W function/the product logarithm function, but I'm not sure how to apply it to get all of the solutions.