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

Maximize $f(x,y,z) = x^4 + y^4 + z^4$ subject to $g(x,y,z) = x^2 + y^2 + z^2 = 1$

it is required that $$\partial_xf = \lambda \partial_xg$$

$$4x^3 = (2x) \lambda \implies x^2 = y^2 = z^2 = \frac{\lambda}{2}$$

$$x^2 + y^2 + z^2 = \frac{ 3 \lambda}{2} = 1 \implies \lambda = \frac{2}{3}$$

$$\therefore x^2 = y^2 = z^2 = \frac{1}{3}$$

and my maximum value is: $\dfrac{3}{81}$

My question is why do I only have one extrema? Also, How do I show this is a max or min.

share|cite|improve this question
Many computations of partials missing. After correct equations are obtained, careful about solving. For example, $4x^3=2\lambda x$ has $x=0$ as a solution. – André Nicolas Jan 16 '13 at 23:14
What you want is to find $\lambda$ such that $\nabla f = \lambda \nabla g.$ More generally, you can show that if $g: \mathbb{R}^n \to \mathbb{R}^{m},$ then a local extremum $a$ of $f: \mathbb{R}^{n} \to \mathbb{R}$ subject to $g = 0$ must satisfy $Df(a) = \lambda_1 Dg_1 (a) + \ldots + \lambda_m Dg_m (a).$ Since $f$ is a functional, $Df(a) = (\nabla f(a))^{T},$ the transpose of the gradient (which is more commonly found in textbooks). – cats Jan 17 '13 at 0:53
In the 3 variable case you can avoid using $\lambda$ at all, since at the critical point(s), the cross-product of the two gradients willo be the zero vector $(0,0,0)$ This usually works quite well I've found. – coffeemath Jan 17 '13 at 2:04
The statement of the constraint has now been changed from what it was originally. So I deleted my answer. – coffeemath Jan 17 '13 at 17:39
up vote 2 down vote accepted

Since $f$ is continuous and the constraint, a sphere, is closed and bounded, you have to have both a maximum and a minimum. So you're right to ask what happened, and @AndréNicolas pointed out in a comment some dropped solutions that lead to the maximum.

An alternate approach is to notice that the problem can rewritten in terms of $u=x^2$, $v=y^2$, $w=z^2$, and is equivalent to maximizing $u^2+v^2+w^2$, which is the square of the distance from the origin, over the equilateral triangle $u+v+w=1$, $0\le u,v,w \le 1$. The extrema occur at the points of the triangle farthest and closest to the origin, which are the vertices and the center respectively.

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.