Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am jammed to problem 11 on page 803 here. Earlier on page 797, it claims that the function changes the least when $\langle \nabla f, \bar{e}\rangle=0$ and it changes the most in the direction of the gradient $\nabla f$.


Suppose $f(x,y,z)=\bar{x}\bar{y}^{2}+\bar{y}\bar{z}^{3}$.

Slowest growth

$\begin{align*} \langle \nabla f, \bar{e}\rangle &= (\bar{x}\bar{y}^{2}+\bar{y}\bar{z}^{3})\cdot \bar{e}_{x}+ (\bar{x}\bar{y}^{2}+\bar{y}\bar{z}^{3})\cdot \bar{e}_{y}+ (\bar{x}\bar{y}^{2}+\bar{y}\bar{z}^{3})\cdot \bar{e}_{z}\\ &= \bar{y}^{2} x +(\bar{x} y^{2}+\bar{z}^{3} y)+(\bar{y}z^{3}) \\&= 0 \end{align*}$

so it is the direction of the slowest growth (according to my book -rule $\langle \nabla f, \bar{e}\rangle =0$, p797)?

Greatest growth direction

$\bar{n}=\max \frac{\nabla f}{\left| f \right|}$ so just I need to find out the inflection points $\nabla^{2} f$ so

$\nabla^{2} f=\begin{pmatrix}\bar{y}^{2}\\ 2\bar{x}\bar{y} \\ 3\bar{y} \bar{z}^{2} \end{pmatrix}=\bar{0},$

I feel I am terribly misunderstanding something.

share|cite|improve this question
up vote 1 down vote accepted

$\DeclareMathOperator{\bR}{\mathbf R}$There does appear to be some confusion. Let $p$ be a point in $\bR^3$ and $f\colon \bR^3 \to \mathbf R^3$ a function. If $u$ is a unit vector in $\bR^3$, then we have the notion of the directional derivative of $f$ at $p$ in the direction of $u$, denoted $D_uf(p)$. Finding the direction of "fastest growth" is finding a $u$ which maximizes this quantity.

If $f$ is differentiable at $p$ then it isn't hard to show that $D_uf(p) = f(p) \cdot u$, where $\nabla f(p)$ is the gradient. Let's make sure that we can compute this: in your example, \[ \nabla f(x, y, z) = (y^2, 2xy + z^3, 3yz^3). \] Using the Schwarz inequality you can conclude that $D_uf(p)$ attains its maximum when $u$ points in the direction of $\nabla f(p)$. The minimum occurs when $u$ points in the opposite direction. If we specialize to $p = (1, 1, 1)$ in the example then $\nabla f(p) = (1, 3, 3)$, and $f$ grows quickest in the direction \[ u = \frac{1}{\sqrt{19}}(1, 3, 3). \]

However, "least change" means minimizing $|D_uf(a)|$. This magnitude is actually zero when $\nabla f(p) \cdot u = 0$, i.e. when $u$ lies in the plane (I should assume that $\nabla f(p) \neq 0$) through the origin with normal vector $\nabla f(p)$. So any unit vector in that plane (and there are infinitely many) will do. Continuing the example, I need to find a $u = (u_1, u_2, u_3)$ such that \[ \begin{pmatrix}1 & 3 & 3\end{pmatrix}\begin{pmatrix}u_1 \\ u_2 \\ u_3\end{pmatrix} = u_1 + 3u_2 + 3u_3 = 0. \] You could find a basis for this null space, or just notice that $(3, -1, 0)$ works and normalize this to get \[ u = \frac{1}{\sqrt{10}}(3, -1, 0). \]

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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