2
$\begingroup$

Reading Nocedal's book on optimization, I came into the concept of local minimizer, which is a well-known concept in numerical optimization. However, I think I am having a rough time trying to come up with an intuitive understanding of the different types of local minimizer. Let us recall the following three definitions:

Def. A point $x^*$ is a weak local minimizer of $f$ if there is a neighborhood $\mathcal{N}$ of $x^*$ such that $f(x^*) \leq f(x)$ for all $x \in \mathcal{N}$.

Def. A point $x^*$ is a strong local minimizer of $f$ if there is a neighborhood $\mathcal{N}$ of $x^*$ such that $f(x^*) < f(x)$ for all $x \in \mathcal{N}$ with $x \neq x^*$.

Def. A point $x^*$ is an isolated local minimizer of $f$ if there is a neighborhood $\mathcal{N}$ of $x^*$ such that $x^*$ is the only local minimizer in $\mathcal{N}$.

Intuitively, a weak local minimizer is a point that achieves the smallest value of $f$ in its neighborhood. A strong local minimizer can be thought off as (to quote Nocedal et al.) "the outright winner in its neighborhood".

But I am failing to see an intuition behind the isolated local minimizer. How is a strong local minimizer not an isolated one, if in the neighborhood, it is the only minimizer?

$\endgroup$
3
$\begingroup$

I'll quote Remark 1.1 from this book:

"Isolated" implies "strict", but the inverse is not true. For example, let $f(x)$ be \begin{equation} f(x) = \begin{cases} 0, & x = 0, \\ x^4 (\cos (1/x) + 2), & x \ne 0. \end{cases} \end{equation} Here, $x^\ast = 0$ is "strict" but not "isolated".

The "strict minimizer" used in the book corresponds to the strong minimizer according to your definition. The point $x^\ast$ is the global minimum, but there are other local minima in every of its neighborhood.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.