Let . The gradient has components and . At a critical point, subtraction gives . The case cannot make either numerator zero, so . The remaining equation is , givingThese are the only stationary points; the denominator is positive everywhere.
At with , differentiating the gradient yields the Hessian matrix . At the positive point it is negative definite, and at the negative point it is positive-definite; in both cases and . Therefore the positive point is a maximum, the negative point is a minimum, and there are no saddle points.
They are global extrema: with , the Cauchy-Schwarz inequality gives , with equality at exactly those two points. This is the extremum bound underlying the level sets of a linear numerator over a radial quadratic.
For a nonzero level , completing the square givesThus the level sets of a linear numerator over a radial quadratic are circles centred on , real only for . The extreme levels shrink to points. The zero contour is the straight line , separating positive and negative heights; near it, small nonzero levels are large circles. Heights tend to zero at infinity.
Circular nonzero contours around the two extrema of a rational height function
. Original contour plot. Blue positive and orange negative circles surround the marked maximum and minimum; the black line is the zero contour. Circle centres lie on . Details. Articles by others on the same topic
There are currently no matching articles.
