3

Say I have a simple 1-D quadratic function $y = a_2x^2 + a_1x + a_0$. I know that if $a$ is positive, there will a unique solution for the turning point, $2$ solutions for any $y$ value above the turning point, and $0$ solutions for $y$ values below the turning point.

Now generalized to arbitrary dimension, i.e. $y = x^TAx - b^Tx + c$, where $x \in \mathbb{R}^n$, the coefficients $A \in \mathbb{R}^{n\times n}$, $b \in \mathbb{R}^n$, and $c$ is a scalar. We can assume $A$ is a symmetric matrix. I am interested in finding the solutions to $x$ given any $y$ value. If my intuitions are correct, there will be infinite solutions for $n \geq 2$ for $y$ values that are above the point where the derivative is $0$. In a 2-D case for example, the solutions will lie on the circle/ellipse formed by a plane cutting the quadratic surface at the height $y$.

How do I compute the length of this circle/ellipse for the 2-D case? Also in higher dimensions, the area (I'm not sure how to describe?) of the solution.

  • For me, determining the set of solutions is disconnected from the computation of the length or the area or the (hyper) volume of this set. What is your motivation for it ? Moreover, if it is an ellipse, its length hasn't a simple mathematical expression. – Jean Marie Sep 28 '23 at 07:17
  • Do you know the semi-axes $a$ and $b$ or the major and minor axes of the ellipse? I can give more info if you do. – John Sep 28 '23 at 08:22
  • For perimeter or surface area, please refer to older posts here, here and here. – Ng Chung Tak Sep 28 '23 at 08:56
  • Pls see usage of quadratic form in older posts here, here and here. – Ng Chung Tak Sep 28 '23 at 09:02
  • @JeanMarie, I'm not sure if it makes sense, but my motivation is to compare the "amount" of possible solutions of the quadratic equation for different $y$ values. Since it's all infinite solutions for any $y$ values above the point of zero derivative, in which case I compare the perimeter instead. Possibly there is a better way to do this? – Tianxun Zhou Sep 28 '23 at 09:17
  • I don't think there exists a mathematical concept making sense for "the amount of solutions" for an infinite set of solutions, and even less that it could be measured by a length... – Jean Marie Sep 28 '23 at 09:23

0 Answers0