It can be shown that
∫ 0 ∞ x 4 sin 2 m + 2 x d x = Q ( 2 m − 1 ) R m π P ( m + 1 ) 2 C S m
for all integers m ≥ 1 , where P , Q , R , S are positive integers with Q , R coprime, and C j is the j th Catalan number C j = j + 1 1 ( j 2 j ) j ≥ 0 .
Give as answer the concatenation P Q R S of the four integers P , Q , R , S . For example, if you think the integral is equal to 7 ( 2 m − 1 ) 3 m π 2 ( m + 1 ) 2 C 5 m , give the answer 2735.
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.
The formula of f m ( y ) can be derived using Laplace transform as well. (See second last comment)
How did you prove that j = 1 ∑ m + 1 ( − 1 ) j j 3 ( m + 1 + j 2 m + 2 ) = 2 ( 2 m + 1 ) ( 2 m − 1 ) ( m + 1 ) ( m + 2 ) ( m + 2 2 m + 2 ) ?
Log in to reply
Well, yes, since f m ( y ) is the Laplace transform of sin 2 m x .
As to your second question, the cute answer is: Mathematica/WA.
To be more helpful, it is a matter of checking the definition that j = 0 ∑ m ( − 1 ) j ( m + j 2 m ) x j = ( m 2 m ) F ( 1 , − m ; m ; x ) where this last hypergeometric function is (because of the − m coefficient) a polynomial of degree m in x . Combining the standard identities: d x d F ( a , b ; c ; x ) = c a b F ( a + 1 , b + 1 ; c + 1 ; x ) and F ( a , b ; c ; 1 ) = Γ ( c − a ) Γ ( c − b ) Γ ( c ) Γ ( c − a − b ) (the last is valid for the ranges of a , b , c that we shall be using), we can calculate A 1 = F ′ ( 1 , − m ; m ; 1 ) A 2 = F ′ ′ ( 1 , − m ; m ; 1 ) A 3 = F ′ ′ ′ ( 1 , − m ; m ; 1 ) and put them together to evaluate A 3 + 3 A 2 + A 1 = j = 1 ∑ m ( − 1 ) j j 3 ( m + j 2 m ) = 2 ( 2 m − 1 ) ( 2 m − 3 ) m 2 The rest is bookkeeping.
Log in to reply
Hmmm, thanks for your detailed explanation (as usual!). It is unfortunate that I don't understand most of your explanation because I don't know how hypergeometric functions actually works, the explanations in Wolfram MathWorld and Wikipedia are too complicated for a commoner like me.
A better question for me to ask right now is: Do you know a good "hypergeometric functions for dummies" book?
On the other hand, is getting "Mathematica" that helpful/ worth it? I looked at the pricing and it doesn't look like it's worthwhile for an amateur hobbyist (such as myself). Thoughts?
Log in to reply
@Pi Han Goh – The 2 F 1 hypergeometric function is given by the series F ( a , b ; c ; x ) = 1 + c a b x + c ( c + 1 ) 2 ! a ( a + 1 ) b ( b + 1 ) x 2 + c ( c + 1 ) ( c + 2 ) 3 ! a ( a + 1 ) ( a + 2 ) b ( b + 1 ) ( b + 2 ) x 3 + ⋯ for ∣ x ∣ < 1 ; in terms of the Pochhammer symbols or Gamma functions F ( a , b ; c ; x ) = j = 0 ∑ ∞ ( c ) j j ! ( a ) j ( b ) j x j = Γ ( a ) Γ ( b ) Γ ( c ) j = 0 ∑ ∞ Γ ( c + j ) j ! Γ ( a + j ) Γ ( b + j ) x j These functions can be analytically continued outside the unit disc. Hypergeometric functions are everywhere in analysis; for a suitable choice of a , b , c , a vast array of functions can be expressed in terms of them.
Whittaker & Watson, "Modern Anaylsis" has a couple of useful chapters. It is old, but so is the study of hypergeometrics, and I think it can be found as a .pdf online! If you just want a summary of results (often all that is necessary), the Digital Library of Mathematical Functions (dlmf.nist.gov) has a very useful section (Chapter 15).
I find Mathematica really useful, not just for calculation and programming, but for constructing diagrams. Learning how to use it has a steep learning curve (I have been to a number of day courses at Wolfram's base in the UK, and still get regularly confused on the programming side). However, I could not afford it were I not a teacher; I enjoy a discount rate, just like students do.
Log in to reply
Well, yes, since f m ( y ) is the Laplace transform of sin 2 m x .
Oh sorry, I just copied it from my Laplace table, I wasn't thinking how it was derived in the first place. Silly me!!
I'll be buying this book as well, estimate time to master it = 5 years
If you just want a summary of results (often all that is necessary), the Digital Library of Mathematical Functions (dlmf.nist.gov) has a very useful section (Chapter 15).
How often does a amateur hobbyist (like myself) need to use this site ? Any difference between the book and the site? I know like less than 0.1% of the stuffs there! Where do I begin to learn them?
Learning how to use it has a steep learning curve (I have been to a number of day courses at Wolfram's base in the UK, and still get regularly confused on the programming side)
Wow! It's even hard for you? Damn! That must be crazy hard to learn then... In the meantime, have you tried SageMath Cloud ? Maybe you're interested!
Log in to reply
@Pi Han Goh – Well, I own a copy of Gradshteyn and Rhyzik, and will doubtless replace it when its spine collapses from overuse, because it contains a huge list of standard results concerning integration. The DLMF is the same. If you just want a result, it is there. If you have time/energy to ask "why"?, then at least you know what you are aiming at! There is more Mathematics out there than any one person can study, and so you have to prioritise which things you are going to accept on authority, and which things you are going to prove for yourself!
As to Mathematica, you (and I) can use it at different levels. For simple calculations, both exact and numeric, it is quite easy to use. What is more demanding is using its programming language, if only because I have yet to find a manual which does not assume a lot more than I know (although there is a lot on online help available). I did use it to draw the animated .gif in "Faster Than Gravity". That is accurate to the model, if at a reduced time rate. I could not have done it otherwise.
Log in to reply
@Mark Hennings – Thank you for your articulated response again. YOU ARE SO WISE!!!!!!!!!!!
@Mark Hennings – What is so special about 2 F 1 hypergeometric function? I read (or just glanced through) some references, and they primarily focuses of 2 F 1 only instead of a F b , where ( a , b ) = ( 2 , 1 ) . Is there a reason why this pair of integers is chosen instead of any other pair?
Log in to reply
@Pi Han Goh – It's a bit simpler than the more general p + 1 F p , and useful enough to be getting on with!
You can use Chu-Vandermonde identity or Snake Oil method to evaluate the binomial sum.
Problem Loading...
Note Loading...
Set Loading...
For any integer m ≥ 0 define f m ( y ) = ∫ 0 ∞ e − x y sin 2 m x d x y > 0 . Integrating by parts twice, we deduce that f m ( y ) = y 2 2 m ( 2 m − 1 ) f m − 1 ( y ) − y 2 4 m 2 f m ( y ) = y 2 + 4 m 2 2 m ( 2 m − 1 ) f m − 1 ( y ) for all m ≥ 1 . Since f 0 ( y ) = y − 1 , it follows that f m ( y ) = y ∏ j = 1 m ( y 2 + 4 j 2 ) ( 2 m ) ! For integers m ≥ n ≥ 1 we have I m , n = = ∫ 0 ∞ x 2 n sin 2 m x d x = ( 2 n − 1 ) ! 1 ∫ 0 ∞ sin 2 m x ( ∫ 0 ∞ y 2 n − 1 e − x y d y ) d x ( 2 n − 1 ) ! 1 ∫ 0 ∞ y 2 n − 1 ( ∫ 0 ∞ e − x y sin 2 m x d x ) d y = ( 2 n − 1 ) ! 1 ∫ 0 ∞ y 2 n − 1 f m ( y ) d y (the reversal of the order of integration is justified by Tonelli's and Fubini's Theorems). Thus I m , n = ( 2 n − 1 ) ! ( 2 m ) ! ∫ 0 ∞ ∏ j = 1 m ( y 2 + 4 j 2 ) y 2 ( n − 1 ) d y m ≥ n ≥ 1 . The integrand can be expanded by partial fractions (in y 2 ), writing ∏ j = 1 m ( y 2 + 4 j 2 ) y 2 ( n − 1 ) = j = 1 ∑ m y 2 + 4 j 2 A j ( m , n ) which makes I m , n = 4 ( 2 n − 1 ) ! π ( 2 m ) ! j = 1 ∑ m j 1 A j ( m , n ) Applying the Cover-Up Rule to determine the coefficients A j ( m , n ) , we deduce that A j ( m , n ) = ( − 1 ) n − j 4 m − n ( m − j ) ! ( m + j ) ! 2 j 2 n and hence I m , n = 4 1 + m − n ( 2 n − 1 ) ! 2 π j = 1 ∑ m ( − 1 ) n − j j 2 n − 1 ( m + j 2 m ) m ≥ n ≥ 1 . For this problem, ∫ 0 ∞ x 4 sin 2 m + 2 x d x = I m + 1 , 2 = = = 4 m 3 ! 2 π j = 1 ∑ m + 1 ( − 1 ) j j 3 ( m + 1 + j 2 m + 2 ) 4 m 3 ! 2 π × 2 ( 2 m + 1 ) ( 2 m − 1 ) ( m + 1 ) ( m + 2 ) ( m + 2 2 m + 2 ) 3 ( 2 m − 1 ) 4 m π ( m + 1 ) ( m 2 m ) = 3 ( 2 m − 1 ) 4 m π ( m + 1 ) 2 C m for any integer m ≥ 1 . This makes the answer 1 3 4 1 .