Skip to lesson

Math · Multivariable Calculus

Chapter 6: Optimization

Critical Points and the Second Derivative Test

Flat spots, and how to tell which kind you found.

Lesson
1
Time
About 24 minutes
0 of 12 done

Step 1: Let's Learn

Read it, or press Listen and follow the words.

A critical point is where both partial derivatives are zero, so the tangent plane is horizontal. Every local extreme is one.

Flat is not enough

A flat spot can be a peak, a pit or a saddle. In one variable there was no third option; in two variables there is.

The discriminant

Compute D = fₓₓfᵧᵧ − (fₓᵧ)². The sign of D and of fₓₓ together decide the classification.

The rules

D > 0 with fₓₓ > 0 is a minimum. D > 0 with fₓₓ < 0 is a maximum. D < 0 is a saddle.

When D is zero

D = 0 tells you nothing. The test declines to answer and other reasoning is needed.

What a saddle is

A saddle is a minimum along one direction and a maximum along another. A mountain pass is the everyday example.

Flat spots

A critical point is where the gradient is zero or undefined. Extrema can only occur there or on the boundary, which reduces an infinite search to a finite list plus a boundary analysis.

Saddle points have no one-variable analogue

A surface can rise in one direction and fall in another at the same flat point. That is a saddle, and it is a genuinely new possibility that two variables introduce.

The second derivative test

The discriminant built from the second partials classifies the point: positive with a positive f_xx gives a minimum, positive with negative gives a maximum, negative gives a saddle, zero is inconclusive.

The boundary must be checked

On a closed bounded region the absolute extrema may lie on the edge, where the gradient need not vanish. Checking the boundary separately is a required step, not an optional refinement.

Step 2: Try It Yourself

Tap and try it out.

The origin here is flat, yet it is neither a peak nor a pit. That is what D < 0 detects.
  • The surfacez = a(x² − y²)
  • Height at the origin0

A saddle curves upward along one axis and downward along the other. The origin is a critical point that is neither a maximum nor a minimum.

The same saddle from above. Contours pulling apart in opposite directions is the signature.

Contours that separate into opposing pairs mean a saddle: uphill one way, downhill the other.

Step 3: Watch an Example

One step at a time.

Watch Hana Classify a Point

Hana has fₓₓ = 4, fᵧᵧ = 9 and fₓᵧ = 0 at a critical point.

  1. Step 1

    She computes D = 4 × 9 − 0², which is 36.

Step 4: Your Turn

Practice makes it stick.

The Test

Problem 1 of 2

fₓₓ = 2, fᵧᵧ = 8 and fₓᵧ = 3. What is D?

The Verdict

Problem 2 of 2

D = −5 at a critical point. Enter 1 for minimum, 2 for maximum, 3 for saddle.

Classify the Flat Spots

1 of 8

fₓₓ = 6, fᵧᵧ = 6, fₓᵧ = 0. What is D?

2 of 8

D = 36 and fₓₓ = 6. Enter 1 for minimum, 2 for maximum, 3 for saddle.

3 of 8

D = 36 and fₓₓ = −6. Enter 1 for minimum, 2 for maximum, 3 for saddle.

4 of 8

fₓₓ = 1, fᵧᵧ = 1, fₓᵧ = 3. What is D?

5 of 8

D = 0. Does the test give an answer? 1 yes, 0 no.

6 of 8

At a critical point, what is the magnitude of the gradient?

7 of 8

Match each case to its classification.

Tap a card on the left to start.

8 of 8

f = x² − y². What is D at the origin, given fₓₓ = 2, fᵧᵧ = −2, fₓᵧ = 0?

Step 5: Quick Check

Show what you know.

Question 1 of 2

fₓₓ = 3, fᵧᵧ = 12, fₓᵧ = 6. What is D?

Question 2 of 2

What is a saddle point?

What You Learned

  • A critical point has both partial derivatives zero, so the tangent plane is horizontal.
  • D = fₓₓfᵧᵧ − (fₓᵧ)² sorts peaks and pits from saddles.
  • D < 0 means a saddle; D = 0 means the test cannot decide.