Here is how you draw the Witch of Agnesi. It's a nice bell-shaped curve that is used as the probability density for the Cauchy distribution. Surprisingly, the region between the curve and its asymptote has no balance point (centroid) [Wiki bit.ly/2TjqYGA] #50FamousCurves
Here is the reason it doesn't have a balance point: The x-coordinate of the centroid of a region between the x axis and the graph of a function y=y(x), A<x<B is x̅ = ∫ x y(x)dx / ∫ y(x)dx, where A and B are the bounds of the definite integrals.
1/4
In the case of the Witch of Agnesi, we have y(x)=a³/(x²+a²) with some fixed a>0 and A=-∞, B=∞. The indefinite integrals are easy to calculate:
∫ y(x)dx = ∫ a³/(x²+a²)dx = a² arctan(x/a)+C
and
∫ x y(x)dx = ∫ x a³/(x²+a²)dx = a³ ln(x²+a²)/2+C'.
2/4
Next we calculate the definite integrals. These are improper integrals because their bounds are at ±∞. The first one is well-defined due to arctan(x)→±π/2 as x→±∞, so the area of the region is finite, namely a²π. It's 4× the area of the circle used in the construction.
3/4
The problem is the second improper integral. It's not well-defined, because its value depends on the way we take the limits of ln(x²+a²) as x→±∞. For example, taking ln(L²+a²)-ln((-L)²+a²) and L→∞ gives 0, but if we take ln((2L)²+a²)-ln((-L)²+a²) and L→∞ we get 2ln(2).
4/4
Fun fact: The name "witch" derives from a mistranslation of the term "averisera" meaning versed sine curve as "avversiera" meaning witch or wife of the devil. #50FamousCurves
Fun fact: The Poisson kernel for the upper half-plane y>0 is the rescaled Witch of Agnesi Pʸ(x)=y/[π(x²+y²)]. This means that heating the edge of the half-plane according to a function f(x) leads to the heat distribution f * Pʸ on the half-plane (*=convolution) . #50FamousCurves
In the animation above f(x)=1 for |x|≤1/2 and f(x)=0 for |x|>1/2. You can see how the convolution with Pʸ gives rise to a smooth heat distribution on the upper half-plane. #50FamousCurves
Fun fact: If you wrap the graph of the tangent function around a cylinder and look at it from the side, you'll see the Witch of Agnesi. #50FamousCurves
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Earlier this year a friend* and I've solved a long-standing problem which, in part, meant finding the eigenvectors of this matrix. In this thread, I'll review our result and bits of 170 years of history
*J.F. van Diejen, Talca
1/n
The title of our paper is "Elliptic Kac–Sylvester Matrix from Difference Lamé Equation" and it was recently published in the mathematical physics journal Annales Henri Poincaré.
Just to "name-drop" some of the characters that will appear in the story: Sylvester (duh), Jacobi, Boltzmann, two Ehrenfests, Schrödinger and Kac (obvs).
(I'll expand the thread over several days so please be patient.)
3/n
"Here's a photo of my boy, Peter. He doesn't yet know what the continuum is, but he doesn't know what fascism is either." - George Szekeres' message to Paul Erdős.
Peter Szekeres was born in Shanghai, where his parents George Szekeres and Esther Klein escaped from Nazi persecution in 1938.
Happy Ending Theorem: any set of five points in the plane in general position has a subset of four points that form the vertices of a convex quadrilateral.
Erdős gave this name to the theorem, because it led to the marriage of Szekeres and Klein
As a run-up to the "Introduction to Integrability" series (see my pinned tweet), I decided to share some interesting bits from the history of integrable systems.
Let's start at the beginning, shall we? So Newton...
#1 It all started with Newton solving the gravitational 2-body problem and deriving Kepler's laws of planetary motion as a result. I would argue that this was possible, because the Kepler problem is (super)integrable. [1/2]
This roughly means that there are many conversed physical quantities like energy, angular momentum, and the Laplace-Runge-Lenz vector. These conservation laws restrict the motion and allow for explicit analytic solutions of otherwise difficult equations. [2/2]