Skip to content
All apps

Local extremum (3D)

About this app

Function f(x,y) as a 3D surface — drag the point, see tangent vectors and find extrema.

The idea

A function of two variables f(x, y) can be drawn as a surface above the x-y plane, with height equal to the function value. At every point of the surface there are two special tangent directions, one along x and one along y. Their slopes are the partial derivatives ∂f/∂x and ∂f/∂y. At a local extremum both vanish at the same time and the tangent plane is horizontal. The necessary condition ∂f/∂x = 0 and ∂f/∂y = 0 generalises f′(x) = 0 from single-variable calculus.

What the app shows

The app shows the smooth, wavy surface f(x, y) = −2·cos(x² + y²)/(x² + y² + 1) + 4. A point P sits on it with two tangent arrows, one along x and one along y. Two sliders for x and y move P across the surface. As soon as both arrows are horizontal, the app reports “Found!”: P is then at a local extremum. The formula and info panels explain the condition.

Try it yourself

Start at x = 1.5 and y = 1.0 and first change only x, then only y. Watch how the matching arrow tilts: the slope in the x direction changes with x but also with y. Then look for a spot where both arrows are flat. The surface is rotationally symmetric because it depends only on x² + y²; rotate the view to see it.

Frequently asked questions

How do you find extrema of a function of two variables?
Compute the partial derivatives ∂f/∂x and ∂f/∂y and set both to zero. The solutions are the critical points. Whether each is a minimum, maximum or saddle point is then decided by the Hessian matrix.
What is a partial derivative intuitively?
The partial derivative ∂f/∂x is the slope of the surface when you move only in the x direction and keep y fixed. Likewise ∂f/∂y measures the slope in the y direction. On the surface these are the slopes of two tangent lines.
Is grad f = 0 enough for an extremum?
No, it is only necessary. Both partial derivatives also vanish at saddle points, which are not extrema. A sufficient condition, for instance via the Hessian, settles the type.
What do I do in the app?
You move point P with the x and y sliders until both tangent arrows are horizontal. The app then reports that a local extremum has been found.

Subject: Analysis (multivariable)

More from Analysis (multivariable)

A VisuApp by heyprof: interactive, free in your browser, no sign-up. What is a VisuApp? · All apps