At a regular constrained extremum, every tangent vector to the level set gives zero directional derivative of . Thus is perpendicular to the tangent plane, and must be a scalar multiple of its nonzero normal vector . The Lagrange multiplier equations are
One solves these four equations for , evaluates at the candidates, and classifies or compares the values as required. Points where , and any boundary or limiting cases of the feasible domain, require separate checks because the regularity argument does not cover them.
Use . The Lagrange multiplier equations are
Taking the dot product with gives , so .
If , the first and third equations give . The middle equation then gives , so . The norm constraint gives and . Thus the hottest points and their temperature are
The coldest points are
For , the equations give , , and hence the other two constrained stationary points are
They are saddle points of the restriction to the sphere: with and as local tangent coordinates there, changes sign in every neighbourhood. The surface is compact and smooth, so the listed extreme temperatures are global. Here “zero gradient on the surface” means zero surface gradient; the ambient gradient at the hottest and coldest points is normal to the sphere, rather than zero.
Introduce the second constraint and a second Lagrange multiplier . At points where the two constraint gradients are linearly independent, solve
These are five scalar equations for . The two constraint normal vectors span the normal plane to the feasible curve. Points where those gradients become dependent must be examined separately; the regular two-multiplier condition need not cover them.

Articles by others on the same topic (0)

There are currently no matching articles.