1
$\begingroup$

Let $f: \mathbb{R}^{n} \to \mathbb{B}$ be continuously differentiable and take the points $x, p \in \mathbb{R}^{n}$. I am familiar with the idea of a directional derivative being given as an inner product of the gradient of a point with another vector in $\mathbb{R}^{n}$. For example, the directional derivative of $f$ in the direction $p$ with respect to the point $x$ can be denoted as $\langle \triangledown f(x), p \rangle$. However, I cannot seem to wrap my mind around the idea of the inner product of the gradient of a point with respect to itself itself. More succinctly, what can we learn about the function $f$ at $p$ given the identity: $$\langle \triangledown f(p), p \rangle$$ This identity can be rewritten as $\langle \triangledown f(p), p \rangle = \sum\limits_{i = 1}^{n}\frac{\partial{f}}{\partial{p_{i}}}p_{i}$, and I'm not quite sure what we can gather from this. I know this is somewhat vague, but if anyone could give me some insight onto the mechanics of what is going on here, I would be very appreciative.

$\endgroup$
2
$\begingroup$

A "direction" is generally given by a unit vector, not just any vector.
If $p$ is not necessarily a unit vector, $\langle \nabla f(x), p \rangle$ would be the rate of change of $f$ as you pass through $x$ with velocity vector $p$.

Now you're looking at the case $x=p$. So your velocity vector is the same as the position vector: the direction is directly away from the origin.

$\endgroup$
  • $\begingroup$ Your explanation is enlightening and much appreciated, sir. $\endgroup$ – Decave Nov 10 '13 at 20:12

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.