relations between Hessian matrix and local extrema
Let be a vector, and let be the Hessian for at a point . Let have continuous
partial derivatives
of first and second order in a neighborhood
of . Let .
If is positive definite (http://planetmath.org/PositiveDefinite), then is a strict local minimum for .
If is a local minimum for , then is positive semidefinite.
If is negative definite (http://planetmath.org/NegativeDefinite), then is a strict local maximum for .
If is a local maximum for , then is negative semidefinite.
If is indefinite, is a nondegenerate saddle point.
If the case when the dimension of is 1 (i.e. ), this reduces to the Second Derivative Test
, which is as follows:
Let the neighborhood of be in the domain for , and let have continuous partial derivatives of first and second order.Let . If , then is a strict local minimum. If , then is a strict local maximum. In the case that , being , is said to be an inflexion point (also called turning point). A typical example is , , , ,, .