Course contentsShow
← Mathematics
Lesson 1868 of 3,12140. Multivariable OptimizationFree lesson

Inconclusive Cases and Further Analysis

Recognize when the second derivative test fails (D = 0) and explore alternative methods for classification.

Inconclusive Cases and Further Analysis

What You'll Learn

You'll learn what to do when the second derivative test can't classify a critical point (when D = 0) and explore alternative strategies to determine its nature.

When the Test Fails

The second derivative test is powerful, but it has a blind spot. Recall that the discriminant D = f_xx · f_yy - (f_xy)² helps classify critical points. When D > 0, you have a local extremum; when D < 0, you have a saddle point.

But what if D = 0? The test is inconclusive—it simply can't tell you what's happening at that critical point. It's like a judge saying "insufficient evidence" rather than "guilty" or "innocent."

Alternative Methods

When D = 0, you need different tools:

1. Examine Cross-Sections

Look at slices of the surface along different paths through the critical point. If all slices curve upward, it's likely a minimum; if all curve downward, a maximum; if some go up and others down, it's a saddle point.

2. Test Nearby Points

Evaluate the function at points surrounding the critical point. Compare these values to the value at the critical point to see if it's a high point, low point, or neither.

3. Use Higher-Order Derivatives

In some cases, examining third or fourth derivatives can reveal behavior that second derivatives miss.

4. Restrict to Level Curves

Analyze the behavior along level curves near the critical point to understand the topology of the surface.

Analogy: Think of D = 0 like standing on a perfectly flat plateau. Looking at the immediate ground (second derivatives) doesn't tell you if you're at a hilltop, valley bottom, or saddle—you need to scout further out.

Key Takeaway: When the second derivative test is inconclusive (D = 0), use alternative methods like examining cross-sections, testing nearby values, or analyzing higher-order behavior to classify the critical point.