Pretty much a circle 3

Calculus Level 5

Let A ( r ) A(r) be the area of the region in the x y xy -plane with x > 0 x>0 and y > 0 y>0 given by the equation Γ ( x ) + Γ ( y ) < r \Gamma(x)+\Gamma(y)<r Let m m be the minimal value of Γ ( x ) \Gamma(x) for x > 0 x>0 , and evaluate the following limit: lim r 2 m + A ( r ) r 2 m \lim_{r\to2m^+}\frac{A(r)}{r-2m} The solution requires numerical computation.


The answer is 7.33183.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Joseph Newton
Aug 23, 2020

This solution is not a complete proof, but shows roughly how the answer is obtained.

Let a a be the value of x x that minimises Γ ( x ) \Gamma(x) , that is Γ ( a ) = m \Gamma(a)=m . As r r decreases towards 2 m 2m (the minimum value of Γ ( x ) + Γ ( y ) \Gamma(x)+\Gamma(y) ), the region Γ ( x ) + Γ ( y ) < r \Gamma(x)+\Gamma(y)<r becomes a smaller and smaller loop around ( x , y ) = ( a , a ) (x,y)=(a,a) , which as it turns out approaches the shape of a circle.

Consider the Taylor series expansion of Γ ( x ) \Gamma(x) around the point x = a x=a : Γ ( x ) = Γ ( a ) + Γ ( a ) ( x a ) + 1 2 Γ ( a ) ( x a ) 2 + \Gamma(x)=\Gamma(a)+\Gamma'(a)(x-a)+\frac12\Gamma''(a)(x-a)^2+\dots Now, Γ ( a ) = m \Gamma(a)=m , and since x = a x=a is a minimum, Γ ( a ) = 0 \Gamma'(a)=0 . We will ignore higher-order terms, and only use a second degree expansion. Intuitively we can do this because as x x approaches a a the terms ( x a ) 3 , ( x a ) 4 , (x-a)^3,(x-a)^4,\dots approach zero much faster than ( x a ) 2 (x-a)^2 . Actually proving that the limit of the area remains the same if we remove these terms is more complex, and not shown here. This leaves Γ ( x ) = m + 1 2 Γ ( a ) ( x a ) 2 \Gamma(x)=m+\frac12\Gamma''(a)(x-a)^2 , so the equation Γ ( x ) + Γ ( y ) < r \Gamma(x)+\Gamma(y)<r becomes m + 1 2 Γ ( a ) ( x a ) 2 + m + 1 2 Γ ( a ) ( y a ) 2 < r m+\frac12\Gamma''(a)(x-a)^2+m+\frac12\Gamma''(a)(y-a)^2<r ( x a ) 2 + ( y a ) 2 < 2 ( r 2 m ) Γ ( a ) \implies(x-a)^2+(y-a)^2<\frac{2(r-2m)}{\Gamma''(a)} This is a circle of radius 2 ( r 2 m ) Γ ( a ) \sqrt{\frac{2(r-2m)}{\Gamma''(a)}} centred at ( a , a ) (a,a) , which has area 2 ( r 2 m ) Γ ( a ) π \frac{2(r-2m)}{\Gamma''(a)}\pi . Hence, the limit becomes lim r 2 m + 1 r 2 m 2 ( r 2 m ) Γ ( a ) π = 2 π Γ ( a ) \lim_{r\to2m^+}\frac1{r-2m}\frac{2(r-2m)}{\Gamma''(a)}\pi=\boxed{\frac{2\pi}{\Gamma''(a)}} Γ ( a ) \Gamma''(a) can be written in terms of the polygamma functions : ψ ( 0 ) ( x ) = d d x ln Γ ( x ) = Γ ( x ) Γ ( x ) \psi^{(0)}(x)=\frac d{dx}\ln\Gamma(x)=\frac{\Gamma'(x)}{\Gamma(x)} ψ ( 1 ) ( x ) = d d x ψ ( 0 ) ( x ) = Γ ( x ) Γ ( x ) ( Γ ( x ) ) 2 ( Γ ( x ) ) 2 = Γ ( x ) Γ ( x ) ( ( ψ ( 0 ) ( x ) ) 2 \psi^{(1)}(x)=\frac d{dx}\psi^{(0)}(x)=\frac{\Gamma''(x)\Gamma(x)-(\Gamma'(x))^2}{(\Gamma(x))^2}=\frac{\Gamma''(x)}{\Gamma(x)}-\left((\psi^{(0)}(x)\right)^2 Γ ( a ) = Γ ( a ) ( ( ψ ( 0 ) ( a ) ) 2 + ψ ( 1 ) ( a ) ) \implies \Gamma''(a)=\Gamma(a)\left(\left(\psi^{(0)}(a)\right)^2+\psi^{(1)}(a)\right) The value of a a can then be found numerically on a program such as Mathematica by looking for the positive root of ψ ( 0 ) ( x ) \psi^{(0)}(x) :

a = x /. FindRoot[PolyGamma[0, x], {x, 2}]

p = Gamma[a] (PolyGamma[0, a]^2 + PolyGamma[1, a])

2 Pi/p

This gives an answer of 7.33183 7.33183 .

This is merely confirmation of Joseph Newton's and must not be taken in any way as criticism.

Yes, Mathematica 12.1.1 gives that answer from that expression. Can be checked on a Raspberry Pi 0 running Raspbian as Wolfram Mathematica is included at no additional cost (Raspbian is free). It is interesting to note that evaluating the limit expression at smaller and smaller values of r r breaks the Mathematica kernel at a value of r r near 0.0000137543 0.0000137543 . This has been reported to Wolfram. At that value of r r , the value of the limit expression is 7.33178 7.33178 , which is close enough to scored as a correct answer by Brilliant's 3 decimal digits standard. The Mathematica kernel issue may be due to issues in the division of the area to be integrated numerically.

Mathematica's confirmation of the second partial derivative:

2 Γ ( t ) t 2 = Γ ( t ) ψ ( 0 ) ( t ) 2 + Γ ( t ) ψ ( 1 ) ( t ) \frac{\partial ^2\Gamma (t)}{\partial t^2}=\Gamma (t) \psi ^{(0)}(t)^2+\Gamma (t) \psi ^{(1)}(t)

A Former Brilliant Member - 9 months, 2 weeks ago

@Joseph Newton

Sorry if this is off topic, but what is the piece of music in your profile picture? It seems like it's in either E major or C# minor. Is it classical?

Krishna Karthik - 9 months ago

Log in to reply

It's from the Rite of Spring, the bit with 6 solo violas

Joseph Newton - 9 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...