The ML inequality is an estimation technique for the upper bound of a contour integral. It tells us that:
where , i.e., the length of . on .
There’s no real perfect bound here: we can go pretty high. But what we do get is a guaranteed upper bound (and not an exact estimate).
Example
Say we want to find an upper bound for:
where is the right-half of a circle where . If we take as the integrand, then we’re looking for:
Since
Then we look at the denominator. We need the denominator term to be an upper bound:
Then we put everything together.