The Hessian Matrix
What You'll Learn
You'll learn how to organize all second-order partial derivatives of a function into a single matrix called the Hessian, which is essential for analyzing critical points.
The Core Idea
When studying functions of two variables, you've already computed second-order partial derivatives: f_xx, f_yy, f_xy, and f_yx. The Hessian matrix simply packages these four derivatives into a 2×2 matrix in a standard way.
For a function f(x, y), the Hessian matrix H is:
H = | f_xx f_xy |
| f_yx f_yy |
Notice the pattern: the first row contains derivatives starting with ∂/∂x, and the second row starts with ∂/∂y. The columns follow the same pattern for the second differentiation.
Why It Matters
Think of the Hessian as a "curvature fingerprint" at a point. Just as the gradient told us about the slope, the Hessian tells us about how the surface bends. This becomes crucial when classifying critical points—is it a hilltop (maximum), valley (minimum), or saddle point?
By Clairaut's Theorem (which you learned earlier), f_xy = f_yx for continuous second derivatives, making the Hessian a symmetric matrix.
Example Construction
For f(x, y) = x³ + xy² - 3y:
- f_x = 3x² + y², so f_xx = 6x
- f_y = 2xy - 3, so f_yy = 2x
- f_xy = 2y (and f_yx = 2y)
The Hessian is:
H = | 6x 2y |
| 2y 2x |
Key Takeaway: The Hessian matrix systematically organizes all second-order partial derivatives into a symmetric 2×2 matrix that captures how a function curves in all directions at a point.