Compute $int_0^{pi} frac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt$
Let $ninmathbb{N}$ and $xin]0,pi[$, I am asked to calculate the following :
$$ I_n = int_0^{pi} dfrac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt$$
From testing on small values of $n$, it seems that this integral is equal to $npicdot cos^n(x)$ but I can't seem to prove it. I tried finding a recurrence formula but didn't succeed.
Here is my working for $n=0$, $n=1$ and $n=2$ :
For $n=0$,
$$ I_0=int_0^{pi}dfrac{cos(x) -cos(t)}{cos(x) - cos(t)}dt = pi$$
For $n=1$,
$$ I_1 = int_0^{pi} dfrac{cos^2(x) -cos^2(t)}{cos(x)-cos(t)}dt=int_0^{pi}cos(x) + sin(t)dt = picdot cos(x)$$
For $n=2$ :
$$ I_2 = int_0^{pi} dfrac{2cos^3(x) - 2cos^3(t) -cos(x) + cos(t)}{cos(x) - cos(t)}dt$$
$$ I_2 = 2int_0^{pi}cos^2(x) +cos(x)cos(t) + cos^2(t) dt - pi$$
$$ I_2 = 2picos^2(x) + int_0^{pi}cos(2t)+1dt - pi$$
$$ I_2 = 2picos^2(x) $$
This is my first post here, please tell me if I did anything wrong. I tried searching this integral on this website without any success.
integration definite-integrals
New contributor
|
show 2 more comments
Let $ninmathbb{N}$ and $xin]0,pi[$, I am asked to calculate the following :
$$ I_n = int_0^{pi} dfrac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt$$
From testing on small values of $n$, it seems that this integral is equal to $npicdot cos^n(x)$ but I can't seem to prove it. I tried finding a recurrence formula but didn't succeed.
Here is my working for $n=0$, $n=1$ and $n=2$ :
For $n=0$,
$$ I_0=int_0^{pi}dfrac{cos(x) -cos(t)}{cos(x) - cos(t)}dt = pi$$
For $n=1$,
$$ I_1 = int_0^{pi} dfrac{cos^2(x) -cos^2(t)}{cos(x)-cos(t)}dt=int_0^{pi}cos(x) + sin(t)dt = picdot cos(x)$$
For $n=2$ :
$$ I_2 = int_0^{pi} dfrac{2cos^3(x) - 2cos^3(t) -cos(x) + cos(t)}{cos(x) - cos(t)}dt$$
$$ I_2 = 2int_0^{pi}cos^2(x) +cos(x)cos(t) + cos^2(t) dt - pi$$
$$ I_2 = 2picos^2(x) + int_0^{pi}cos(2t)+1dt - pi$$
$$ I_2 = 2picos^2(x) $$
This is my first post here, please tell me if I did anything wrong. I tried searching this integral on this website without any success.
integration definite-integrals
New contributor
2
Hello, welcome to MSE. This isn't a bad post as a new contributor, but you (and us) would benefit by adding some proofs for your 'testing' - show us your working for $n=1,2$. Good job on using MathJax!
– TheSimpliFire
Jan 4 at 16:56
@TheSimpliFire Added my working for n = 1, 2, thanks for the tip and the welcome., I need to calculate this integral to complete a problem from by problemset, which consists in calculating $int_0^{pi} dfrac{cos(nx) - cos(nt)}{cos(x) - cos(t)}dt$
– aleph0
Jan 4 at 17:13
Very nice solution (+1).
– TheSimpliFire
Jan 4 at 17:14
@aleph0 So you need to calculate$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$?
– Frank W.
Jan 4 at 17:16
1
Are you sure that the result is correct?
– Zacky
Jan 4 at 17:41
|
show 2 more comments
Let $ninmathbb{N}$ and $xin]0,pi[$, I am asked to calculate the following :
$$ I_n = int_0^{pi} dfrac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt$$
From testing on small values of $n$, it seems that this integral is equal to $npicdot cos^n(x)$ but I can't seem to prove it. I tried finding a recurrence formula but didn't succeed.
Here is my working for $n=0$, $n=1$ and $n=2$ :
For $n=0$,
$$ I_0=int_0^{pi}dfrac{cos(x) -cos(t)}{cos(x) - cos(t)}dt = pi$$
For $n=1$,
$$ I_1 = int_0^{pi} dfrac{cos^2(x) -cos^2(t)}{cos(x)-cos(t)}dt=int_0^{pi}cos(x) + sin(t)dt = picdot cos(x)$$
For $n=2$ :
$$ I_2 = int_0^{pi} dfrac{2cos^3(x) - 2cos^3(t) -cos(x) + cos(t)}{cos(x) - cos(t)}dt$$
$$ I_2 = 2int_0^{pi}cos^2(x) +cos(x)cos(t) + cos^2(t) dt - pi$$
$$ I_2 = 2picos^2(x) + int_0^{pi}cos(2t)+1dt - pi$$
$$ I_2 = 2picos^2(x) $$
This is my first post here, please tell me if I did anything wrong. I tried searching this integral on this website without any success.
integration definite-integrals
New contributor
Let $ninmathbb{N}$ and $xin]0,pi[$, I am asked to calculate the following :
$$ I_n = int_0^{pi} dfrac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt$$
From testing on small values of $n$, it seems that this integral is equal to $npicdot cos^n(x)$ but I can't seem to prove it. I tried finding a recurrence formula but didn't succeed.
Here is my working for $n=0$, $n=1$ and $n=2$ :
For $n=0$,
$$ I_0=int_0^{pi}dfrac{cos(x) -cos(t)}{cos(x) - cos(t)}dt = pi$$
For $n=1$,
$$ I_1 = int_0^{pi} dfrac{cos^2(x) -cos^2(t)}{cos(x)-cos(t)}dt=int_0^{pi}cos(x) + sin(t)dt = picdot cos(x)$$
For $n=2$ :
$$ I_2 = int_0^{pi} dfrac{2cos^3(x) - 2cos^3(t) -cos(x) + cos(t)}{cos(x) - cos(t)}dt$$
$$ I_2 = 2int_0^{pi}cos^2(x) +cos(x)cos(t) + cos^2(t) dt - pi$$
$$ I_2 = 2picos^2(x) + int_0^{pi}cos(2t)+1dt - pi$$
$$ I_2 = 2picos^2(x) $$
This is my first post here, please tell me if I did anything wrong. I tried searching this integral on this website without any success.
integration definite-integrals
integration definite-integrals
New contributor
New contributor
edited Jan 4 at 17:38
Did
246k23221456
246k23221456
New contributor
asked Jan 4 at 16:53
aleph0aleph0
585
585
New contributor
New contributor
2
Hello, welcome to MSE. This isn't a bad post as a new contributor, but you (and us) would benefit by adding some proofs for your 'testing' - show us your working for $n=1,2$. Good job on using MathJax!
– TheSimpliFire
Jan 4 at 16:56
@TheSimpliFire Added my working for n = 1, 2, thanks for the tip and the welcome., I need to calculate this integral to complete a problem from by problemset, which consists in calculating $int_0^{pi} dfrac{cos(nx) - cos(nt)}{cos(x) - cos(t)}dt$
– aleph0
Jan 4 at 17:13
Very nice solution (+1).
– TheSimpliFire
Jan 4 at 17:14
@aleph0 So you need to calculate$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$?
– Frank W.
Jan 4 at 17:16
1
Are you sure that the result is correct?
– Zacky
Jan 4 at 17:41
|
show 2 more comments
2
Hello, welcome to MSE. This isn't a bad post as a new contributor, but you (and us) would benefit by adding some proofs for your 'testing' - show us your working for $n=1,2$. Good job on using MathJax!
– TheSimpliFire
Jan 4 at 16:56
@TheSimpliFire Added my working for n = 1, 2, thanks for the tip and the welcome., I need to calculate this integral to complete a problem from by problemset, which consists in calculating $int_0^{pi} dfrac{cos(nx) - cos(nt)}{cos(x) - cos(t)}dt$
– aleph0
Jan 4 at 17:13
Very nice solution (+1).
– TheSimpliFire
Jan 4 at 17:14
@aleph0 So you need to calculate$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$?
– Frank W.
Jan 4 at 17:16
1
Are you sure that the result is correct?
– Zacky
Jan 4 at 17:41
2
2
Hello, welcome to MSE. This isn't a bad post as a new contributor, but you (and us) would benefit by adding some proofs for your 'testing' - show us your working for $n=1,2$. Good job on using MathJax!
– TheSimpliFire
Jan 4 at 16:56
Hello, welcome to MSE. This isn't a bad post as a new contributor, but you (and us) would benefit by adding some proofs for your 'testing' - show us your working for $n=1,2$. Good job on using MathJax!
– TheSimpliFire
Jan 4 at 16:56
@TheSimpliFire Added my working for n = 1, 2, thanks for the tip and the welcome., I need to calculate this integral to complete a problem from by problemset, which consists in calculating $int_0^{pi} dfrac{cos(nx) - cos(nt)}{cos(x) - cos(t)}dt$
– aleph0
Jan 4 at 17:13
@TheSimpliFire Added my working for n = 1, 2, thanks for the tip and the welcome., I need to calculate this integral to complete a problem from by problemset, which consists in calculating $int_0^{pi} dfrac{cos(nx) - cos(nt)}{cos(x) - cos(t)}dt$
– aleph0
Jan 4 at 17:13
Very nice solution (+1).
– TheSimpliFire
Jan 4 at 17:14
Very nice solution (+1).
– TheSimpliFire
Jan 4 at 17:14
@aleph0 So you need to calculate$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$?
– Frank W.
Jan 4 at 17:16
@aleph0 So you need to calculate$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$?
– Frank W.
Jan 4 at 17:16
1
1
Are you sure that the result is correct?
– Zacky
Jan 4 at 17:41
Are you sure that the result is correct?
– Zacky
Jan 4 at 17:41
|
show 2 more comments
5 Answers
5
active
oldest
votes
Here's a solution that only rests on the following simple trigonometric identity:
$$cos(a+b)+cos(a-b)=2cos(a)cos(b)tag{1}$$
We'll get back to it later, but for now, notice that
$$begin{split}
I_n(x)&=int_0^{pi} frac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt\
&=int_0^{pi}frac{[cos(nx)-cos(nt)]cos(x) + cos(nt)[cos(x)-cos(t)]}{cos(x) -cos(t)}dt\
&=cos(x)int_0^{pi}frac{cos(nx)-cos(nt)}{cos(x) -cos(t)}dt+int_0^picos(nt)dt
end{split}$$
In other words,
$$I_n(x)=cos(x)J_n(x)+pidelta_{n=0}tag{2}$$
where we define $$J_n(x)=int_0^pi frac{cos(nx)-cos(nt)}{cos(x)-cos(t)}dt$$
and the Kronecker symbol $delta_{n=0}$, which is equal $0$, unless $n=0$, in which case it's equal to $1$.
Now, let's go back to (1). Plugging $a=nx$ and $b=x$ into that identity implies that
$$cos((n+1)x)+cos((n-1)x)=2cos x cos(nx)$$
Subtracting the same equation with $t$ to this one yields
$$
begin{split}
cos((n+1)x)-cos((n+1)t) \
+cos((n-1)x)-cos((n-1)t)=\
2cos x cos(nx)-2cos(t)cos(nt)
end{split}$$
Dividing by $cos(x)-cos(t)$, and integrating over $[0,pi]$ leads to
$$J_{n+1}(x)+J_{n-1}(x)=2I_n(x)tag{3}$$
Finally, combining [2] and [3] gets us, for $ngeq 0$,
$$J_{n+2}(x)-2cos(x)J_{n+1}(x)+J_{n}(x)=0$$
The solution to this second-order recurrence relation is
$$J_n(x)=alpha e^{inx}+beta e^{-inx}$$
Since, $J_0=0$ and $J_1=pi$,
$$J_n(x)=frac {pi sin(nx)}{sin x}$$
and $$I_n(x)=picos(x)frac{sin(nx)}{sin(x)} mbox{ for } ngeq 1 mbox{, and }I_0=pi$$
Thanks a lot for that solution !
– aleph0
Jan 4 at 22:41
You're welcome!
– Stefan Lafon
Jan 4 at 23:05
I think there is a small mistake: We have $I_n = (J_{n+1} color{red}{+} J_{n-1})/2$, which turns the final result into $I_n (x) = pi cos(x) frac{sin(n x)}{sin(x)}$ .
– ComplexYetTrivial
Jan 5 at 8:27
Thanks for catching this! I updated the answer.
– Stefan Lafon
Jan 6 at 0:40
add a comment |
A nice way to evaluate generalized integrals is to consider them as the coefficients of an infinite series. Therefore, the coefficient of the $n$th term is simply the integral under question. Before we begin though, there is one identity to note:
$$sumlimits_{ngeq0}z^ncos nx=frac {1-zcos x}{z^2-2zcos x+1}$$
Proof: Rewrite $cos nx$ as the real part of $e^{nix}$. Using the infinite geometric sequences, we get that$$sumlimits_{ngeq0}left(ze^{ix}right)^n=frac 1{1-ze^{ix}}$$Now, take the real part of both sides. Clearly, the left - hand side becomes $z^ncos nx$. Meanwhile, the right - hand side becomes, through some clever rationalization$$begin{align*}operatorname{Re}left[frac 1{1-ze^{ix}}right] & =operatorname{Re}left[frac 1{1-zcos x-zisin x}right]\ & =operatorname{Re}left[frac {1-zcos x+zisin x}{(1-zcos x)^2+z^2sin^2x}right]\ & =frac {1-zcos x}{z^2-2zcos x+1}end{align*}$$completing the proof.
With that in mind, we are ready to begin. Since the OP has stated in the comments that he is trying to evaluate the integral$$I_n=intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$I will show a way to evaluate it in this answer. To wit, denote the generating function of the integral as $G(z)$
$$G(z)=sumlimits_{ngeq0}I_nz^n$$
And remember that the coefficient of $z^n$ simply gives $I_n$. Interchange the sum and the integral, and using the identity we've derived above, get
$$begin{align*}G(z) & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}sumlimits_{ngeq0}z^nbiggr[cos nx-cos ntbiggr]\ & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}left[frac {1-zcos x}{z^2-2zcos x+1}-frac {1-zcos t}{z^2-2zcos t+1}right]end{align*}$$
Combining the two fractions, and recalling that any terms in $z$ are constants, the function becomes
$$G(z)=frac {z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{pi}frac {mathrm dt}{z^2-2zcos t+1}$$
The remaining integral can be easily evaluated using a Weierstrass substitution. Substitute $w=tanleft(tfrac t2right)$ so that
$$begin{array}{|c|c|c|}hline w=tanleft(dfrac t2right) & mathrm dt=dfrac {2,mathrm dw}{1+w^2} & cos t=dfrac {1-w^2}{1+w^2}\hlineend{array}$$
The remaining rational function can be evaluated in an elementary fashion
$$begin{align*}G(z) & =frac {2z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{infty}frac {mathrm dw}{w^2(1+z)^2+(1-z)^2}\ & =frac {2z}{z^2-2zcos x+1}arctanleft(frac {1+z}{1-z}wright),Biggrrvert_0^{infty}\ & =frac {pi z}{z^2-2zcos x+1}end{align*}$$
From the second line, it's important to observe that the argument of the inverse tangent will remain positive if and only if $|z|<1$. When $|z|<1$, then the denominator is positive, as well as the numerator. Therefore, $tfrac {1+z}{1-z}>0$. However, if $|z|>1$, then the argument is less than zero and there is an extra negative sign. For the purpose of this question, we'll consider when $|z|<1$.
Now all we have to do is find the coefficient of $z^n$. There is a nice and convenient way to do this by using
$$2cos x=e^{ix}+e^{-ix}$$
Factoring the denominator by grouping gives
$$begin{align*}frac z{z^2-2zcos x+1} & =frac z{(1-ze^{ix})(1-ze^{-ix})}\ & =zsumlimits_{kgeq0}z^k e^{kix}sumlimits_{lgeq0}z^l e^{-lix}end{align*}$$
Now observe what happens when we expand the products together$$begin{multline}(1+ze^{ix}+z^2e^{2ix}+cdots)(1+ze^{-ix}+z^2e^{-ix}+cdots)\=1+z(e^{ix}+e^{-ix})+z^2(e^{2ix}+1+e^{-2ix})+cdotsend{multline}$$
The sum within the parenthesis seems to start off at the index of the $n$th term and decrease by a factor of two! Using this, it's possible to rewrite the coefficients conveniently as
$$a_k=sumlimits_{m=0}^ke^{(k-2m)ix}=frac {sin x(k+1)}{sin x}$$
Hence$$frac {pi z}{z^2-2zcos x+1}=pisumlimits_{kgeq1}frac {sin xk}{sin x}z^k$$
And setting $k$ as $n$ gives the term $z^n$. Therefore, our integral is simply$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}color{blue}{=frac {pisin xn}{sin x}}$$
@StefanLafon I'm having a little bit of trouble with that; more than I thought initially. I'm trying to mold the infinite sum into$$frac 1{z^2-2zcos x+1}$$on the right - hand side and a $z^n$ term on the left. It's not going very well I must say
– Frank W.
Jan 4 at 18:00
Impressive. To find the coefficients, notice that $frac {2isin x} {z^2-2xcos(x)+1} = frac 1 {z+e^{-ix}}-frac 1 {z+e^{ix}}$.
– Stefan Lafon
Jan 4 at 18:01
Correct the sign mistake in previous comment and show the $n$th integral is $$I_n=pifrac{sin nx}{sin x}$$
– Did
Jan 4 at 18:05
Thanks a lot for your time and effort. This is a very interesting solution but the use of generating functions seems a bit overkill and I don't understand all the details, as I have not yet studied generating functions in details. This problem is supposed to be a medium difficulty problem of first year at university . The tip below the problem suggested to calculate the integral in the question.
– aleph0
Jan 4 at 18:10
1
@Zacky Oh okay. I missed it.
– Frank W.
Jan 4 at 18:55
|
show 3 more comments
Completing Frank's solution:
$$ [z^n]frac{pi z}{z^2-2zcos x+1} = frac{pi}{2}[z^{n}]left(frac{1}{z-e^{ix}}+frac{1}{z-e^{-ix}}right) $$
equals, by geometric series,
$$ frac{pi}{2}left(-e^{-(n+1)ix}-e^{(n+1)ix}right)=-picos((n+1)x). $$
add a comment |
I will do the integral for $n=3$, unfortunately this will disprove your given result. (unless I did a mistake)
$$I_3= int_0^pi frac{cos(3x)cos x -cos(3t)cos t}{cos x-cos t}dt$$
Since $cos(3 y)=4cos^3 y -3cos y$ we have: $$cos(3x)cos x -cos (3t)cos t=4(cos^4 x-cos^4t)-3(cos^2 x-cos^2 t)$$
$$=4(cos x-cos t)(cos x+cos t) (cos^2 x+cos^2t)-3(cos x-cos t)(cos x+cos t)$$
$$Rightarrow I_3=int_0^pi (cos x+cos t)(4(cos ^2 x+cos^2 t)-3)dt$$
$$overset{pi-tto t}=int_0^pi (cos x-cos t)(4(cos^2 x+cos^2 t)-3)dt$$
$$Rightarrow 2I_3=2cos xint_0^pi (4(cos^2 x+cos^2 t)-3)$$
$$Rightarrow I_3= 4pi cos^3 x +cos x underbrace{int_0^pi cos^2 tdt}_{=frac{pi}{2}}-3pi cos x=2pi(cos x+2cos(3x))$$
For $n=4$ we have: $$cos(4x)=8cos^4 x-8cos^2 x+1$$
Denoting $cx=cos x$ and $ct=cos t,$ we get the integrand to be:
$$frac{8(cx^5-ct^5)-8(cx^3-ct^3)+(cx-ct)}{cx-ct}$$
And using the fact that: $$a^5-b^5=(a-b)(a^4+a^3b+a^2b^2+ab^3+b^4$$
$$Rightarrow I_4=8int_0^pi(cx^4+cx^3 ct+cx^2 ct^2 +cx ct^3+ct^4)dt$$
$$-8int_0^pi (cx^2+cxct+ct^2) dt +int_0^pi dt$$
We have: $$int_0^pi ct dt=int_0^pi ct^3 dt =0$$
$$int_0^pi ct^2 dt= frac{pi}{2}, int_0^pi ct^4 dt=frac{3pi}{8}$$
$$Rightarrow I_4=(8pi cx^4 +4pi cx^2 +3pi )-(8pi cx^2 +4pi)+pi$$
$$=8pi cx^4 -4pi cx^2 =4pi cos^2 x cos(2x)$$
1
I'm getting a slightly different answer too than what the OP conjectured.
– Frank W.
Jan 4 at 18:00
1
It seems like it only works for $n=1$ and $n=2$ until now. Note that even $n=0$ fails.
– Zacky
Jan 4 at 18:01
Thanks for doing that. I guess my conjecture was wrong...
– aleph0
Jan 4 at 18:08
add a comment |
An alternative solution to the problem:
For $n in mathbb{N}$ and $x in (0,pi)$ define
$$J_n (x) equiv int limits_0^pi frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t , . $$
We can use the identities ($(2)$ follows from the geometric progression formula)
begin{align}
cos(xi) - cos(tau) &= - 2 sin left(frac{xi + tau}{2}right) sin left(frac{xi - tau}{2}right) , , , xi,tau in mathbb{R} , , tag{1} \
frac{sin(n y)}{sin(y)} &= mathrm{e}^{-mathrm{i}(n-1)y} sum limits_{k=0}^{n-1} mathrm{e}^{2mathrm{i} k y} , , , n in mathbb{N} , , , y in mathbb{R} , , tag{2} \
int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t &= 2 pi delta_{k,l} , , , k,l in mathbb{Z} , , tag{3}
end{align}
to compute
begin{align}
J_n (x) &= frac{1}{2} int limits_0^{2pi} frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t stackrel{(1)}{=} frac{1}{2} int limits_0^{2pi} frac{sin left(nfrac{x+t}{2}right)}{sin left(frac{x+t}{2}right)} frac{sin left(nfrac{x-t}{2}right)}{sin left(frac{x-t}{2}right)} , mathrm{d} t \
&stackrel{(2)}{=} frac{1}{2} mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k,l=0}^{n-1} mathrm{e}^{mathrm{i} (k+l) x} int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t stackrel{(3)}{=} pi mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k=0}^{n-1} mathrm{e}^{2 mathrm{i} k x} \
&stackrel{(2)}{=} pi frac{sin(nx)}{sin(x)} , .
end{align}
This result directly leads to
begin{align}
I_n(x) &equiv int limits_0^pi frac{cos(n x) cos(x) - cos(n t) cos(t)}{cos(x) - cos(t)} , mathrm{d} t = int limits_0^pi left[cos(x)frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} + cos(n t)right], mathrm{d} t \
&= cos(x) J_n(x) + 0 = pi cos(x) frac{sin(nx)}{sin(x)} , .
end{align}
Great solution ! Thank you.
– aleph0
Jan 5 at 11:48
add a comment |
Your Answer
StackExchange.ifUsing("editor", function () {
return StackExchange.using("mathjaxEditing", function () {
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
});
});
}, "mathjax-editing");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "69"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
noCode: true, onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
aleph0 is a new contributor. Be nice, and check out our Code of Conduct.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3061827%2fcompute-int-0-pi-frac-cosnx-cosx-cosnt-cost-cosx-cost%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
5 Answers
5
active
oldest
votes
5 Answers
5
active
oldest
votes
active
oldest
votes
active
oldest
votes
Here's a solution that only rests on the following simple trigonometric identity:
$$cos(a+b)+cos(a-b)=2cos(a)cos(b)tag{1}$$
We'll get back to it later, but for now, notice that
$$begin{split}
I_n(x)&=int_0^{pi} frac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt\
&=int_0^{pi}frac{[cos(nx)-cos(nt)]cos(x) + cos(nt)[cos(x)-cos(t)]}{cos(x) -cos(t)}dt\
&=cos(x)int_0^{pi}frac{cos(nx)-cos(nt)}{cos(x) -cos(t)}dt+int_0^picos(nt)dt
end{split}$$
In other words,
$$I_n(x)=cos(x)J_n(x)+pidelta_{n=0}tag{2}$$
where we define $$J_n(x)=int_0^pi frac{cos(nx)-cos(nt)}{cos(x)-cos(t)}dt$$
and the Kronecker symbol $delta_{n=0}$, which is equal $0$, unless $n=0$, in which case it's equal to $1$.
Now, let's go back to (1). Plugging $a=nx$ and $b=x$ into that identity implies that
$$cos((n+1)x)+cos((n-1)x)=2cos x cos(nx)$$
Subtracting the same equation with $t$ to this one yields
$$
begin{split}
cos((n+1)x)-cos((n+1)t) \
+cos((n-1)x)-cos((n-1)t)=\
2cos x cos(nx)-2cos(t)cos(nt)
end{split}$$
Dividing by $cos(x)-cos(t)$, and integrating over $[0,pi]$ leads to
$$J_{n+1}(x)+J_{n-1}(x)=2I_n(x)tag{3}$$
Finally, combining [2] and [3] gets us, for $ngeq 0$,
$$J_{n+2}(x)-2cos(x)J_{n+1}(x)+J_{n}(x)=0$$
The solution to this second-order recurrence relation is
$$J_n(x)=alpha e^{inx}+beta e^{-inx}$$
Since, $J_0=0$ and $J_1=pi$,
$$J_n(x)=frac {pi sin(nx)}{sin x}$$
and $$I_n(x)=picos(x)frac{sin(nx)}{sin(x)} mbox{ for } ngeq 1 mbox{, and }I_0=pi$$
Thanks a lot for that solution !
– aleph0
Jan 4 at 22:41
You're welcome!
– Stefan Lafon
Jan 4 at 23:05
I think there is a small mistake: We have $I_n = (J_{n+1} color{red}{+} J_{n-1})/2$, which turns the final result into $I_n (x) = pi cos(x) frac{sin(n x)}{sin(x)}$ .
– ComplexYetTrivial
Jan 5 at 8:27
Thanks for catching this! I updated the answer.
– Stefan Lafon
Jan 6 at 0:40
add a comment |
Here's a solution that only rests on the following simple trigonometric identity:
$$cos(a+b)+cos(a-b)=2cos(a)cos(b)tag{1}$$
We'll get back to it later, but for now, notice that
$$begin{split}
I_n(x)&=int_0^{pi} frac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt\
&=int_0^{pi}frac{[cos(nx)-cos(nt)]cos(x) + cos(nt)[cos(x)-cos(t)]}{cos(x) -cos(t)}dt\
&=cos(x)int_0^{pi}frac{cos(nx)-cos(nt)}{cos(x) -cos(t)}dt+int_0^picos(nt)dt
end{split}$$
In other words,
$$I_n(x)=cos(x)J_n(x)+pidelta_{n=0}tag{2}$$
where we define $$J_n(x)=int_0^pi frac{cos(nx)-cos(nt)}{cos(x)-cos(t)}dt$$
and the Kronecker symbol $delta_{n=0}$, which is equal $0$, unless $n=0$, in which case it's equal to $1$.
Now, let's go back to (1). Plugging $a=nx$ and $b=x$ into that identity implies that
$$cos((n+1)x)+cos((n-1)x)=2cos x cos(nx)$$
Subtracting the same equation with $t$ to this one yields
$$
begin{split}
cos((n+1)x)-cos((n+1)t) \
+cos((n-1)x)-cos((n-1)t)=\
2cos x cos(nx)-2cos(t)cos(nt)
end{split}$$
Dividing by $cos(x)-cos(t)$, and integrating over $[0,pi]$ leads to
$$J_{n+1}(x)+J_{n-1}(x)=2I_n(x)tag{3}$$
Finally, combining [2] and [3] gets us, for $ngeq 0$,
$$J_{n+2}(x)-2cos(x)J_{n+1}(x)+J_{n}(x)=0$$
The solution to this second-order recurrence relation is
$$J_n(x)=alpha e^{inx}+beta e^{-inx}$$
Since, $J_0=0$ and $J_1=pi$,
$$J_n(x)=frac {pi sin(nx)}{sin x}$$
and $$I_n(x)=picos(x)frac{sin(nx)}{sin(x)} mbox{ for } ngeq 1 mbox{, and }I_0=pi$$
Thanks a lot for that solution !
– aleph0
Jan 4 at 22:41
You're welcome!
– Stefan Lafon
Jan 4 at 23:05
I think there is a small mistake: We have $I_n = (J_{n+1} color{red}{+} J_{n-1})/2$, which turns the final result into $I_n (x) = pi cos(x) frac{sin(n x)}{sin(x)}$ .
– ComplexYetTrivial
Jan 5 at 8:27
Thanks for catching this! I updated the answer.
– Stefan Lafon
Jan 6 at 0:40
add a comment |
Here's a solution that only rests on the following simple trigonometric identity:
$$cos(a+b)+cos(a-b)=2cos(a)cos(b)tag{1}$$
We'll get back to it later, but for now, notice that
$$begin{split}
I_n(x)&=int_0^{pi} frac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt\
&=int_0^{pi}frac{[cos(nx)-cos(nt)]cos(x) + cos(nt)[cos(x)-cos(t)]}{cos(x) -cos(t)}dt\
&=cos(x)int_0^{pi}frac{cos(nx)-cos(nt)}{cos(x) -cos(t)}dt+int_0^picos(nt)dt
end{split}$$
In other words,
$$I_n(x)=cos(x)J_n(x)+pidelta_{n=0}tag{2}$$
where we define $$J_n(x)=int_0^pi frac{cos(nx)-cos(nt)}{cos(x)-cos(t)}dt$$
and the Kronecker symbol $delta_{n=0}$, which is equal $0$, unless $n=0$, in which case it's equal to $1$.
Now, let's go back to (1). Plugging $a=nx$ and $b=x$ into that identity implies that
$$cos((n+1)x)+cos((n-1)x)=2cos x cos(nx)$$
Subtracting the same equation with $t$ to this one yields
$$
begin{split}
cos((n+1)x)-cos((n+1)t) \
+cos((n-1)x)-cos((n-1)t)=\
2cos x cos(nx)-2cos(t)cos(nt)
end{split}$$
Dividing by $cos(x)-cos(t)$, and integrating over $[0,pi]$ leads to
$$J_{n+1}(x)+J_{n-1}(x)=2I_n(x)tag{3}$$
Finally, combining [2] and [3] gets us, for $ngeq 0$,
$$J_{n+2}(x)-2cos(x)J_{n+1}(x)+J_{n}(x)=0$$
The solution to this second-order recurrence relation is
$$J_n(x)=alpha e^{inx}+beta e^{-inx}$$
Since, $J_0=0$ and $J_1=pi$,
$$J_n(x)=frac {pi sin(nx)}{sin x}$$
and $$I_n(x)=picos(x)frac{sin(nx)}{sin(x)} mbox{ for } ngeq 1 mbox{, and }I_0=pi$$
Here's a solution that only rests on the following simple trigonometric identity:
$$cos(a+b)+cos(a-b)=2cos(a)cos(b)tag{1}$$
We'll get back to it later, but for now, notice that
$$begin{split}
I_n(x)&=int_0^{pi} frac{cos(nx)cos(x) - cos(nt)cos(t)}{cos(x) -cos(t)}dt\
&=int_0^{pi}frac{[cos(nx)-cos(nt)]cos(x) + cos(nt)[cos(x)-cos(t)]}{cos(x) -cos(t)}dt\
&=cos(x)int_0^{pi}frac{cos(nx)-cos(nt)}{cos(x) -cos(t)}dt+int_0^picos(nt)dt
end{split}$$
In other words,
$$I_n(x)=cos(x)J_n(x)+pidelta_{n=0}tag{2}$$
where we define $$J_n(x)=int_0^pi frac{cos(nx)-cos(nt)}{cos(x)-cos(t)}dt$$
and the Kronecker symbol $delta_{n=0}$, which is equal $0$, unless $n=0$, in which case it's equal to $1$.
Now, let's go back to (1). Plugging $a=nx$ and $b=x$ into that identity implies that
$$cos((n+1)x)+cos((n-1)x)=2cos x cos(nx)$$
Subtracting the same equation with $t$ to this one yields
$$
begin{split}
cos((n+1)x)-cos((n+1)t) \
+cos((n-1)x)-cos((n-1)t)=\
2cos x cos(nx)-2cos(t)cos(nt)
end{split}$$
Dividing by $cos(x)-cos(t)$, and integrating over $[0,pi]$ leads to
$$J_{n+1}(x)+J_{n-1}(x)=2I_n(x)tag{3}$$
Finally, combining [2] and [3] gets us, for $ngeq 0$,
$$J_{n+2}(x)-2cos(x)J_{n+1}(x)+J_{n}(x)=0$$
The solution to this second-order recurrence relation is
$$J_n(x)=alpha e^{inx}+beta e^{-inx}$$
Since, $J_0=0$ and $J_1=pi$,
$$J_n(x)=frac {pi sin(nx)}{sin x}$$
and $$I_n(x)=picos(x)frac{sin(nx)}{sin(x)} mbox{ for } ngeq 1 mbox{, and }I_0=pi$$
edited Jan 6 at 5:16
answered Jan 4 at 22:07
Stefan LafonStefan Lafon
98616
98616
Thanks a lot for that solution !
– aleph0
Jan 4 at 22:41
You're welcome!
– Stefan Lafon
Jan 4 at 23:05
I think there is a small mistake: We have $I_n = (J_{n+1} color{red}{+} J_{n-1})/2$, which turns the final result into $I_n (x) = pi cos(x) frac{sin(n x)}{sin(x)}$ .
– ComplexYetTrivial
Jan 5 at 8:27
Thanks for catching this! I updated the answer.
– Stefan Lafon
Jan 6 at 0:40
add a comment |
Thanks a lot for that solution !
– aleph0
Jan 4 at 22:41
You're welcome!
– Stefan Lafon
Jan 4 at 23:05
I think there is a small mistake: We have $I_n = (J_{n+1} color{red}{+} J_{n-1})/2$, which turns the final result into $I_n (x) = pi cos(x) frac{sin(n x)}{sin(x)}$ .
– ComplexYetTrivial
Jan 5 at 8:27
Thanks for catching this! I updated the answer.
– Stefan Lafon
Jan 6 at 0:40
Thanks a lot for that solution !
– aleph0
Jan 4 at 22:41
Thanks a lot for that solution !
– aleph0
Jan 4 at 22:41
You're welcome!
– Stefan Lafon
Jan 4 at 23:05
You're welcome!
– Stefan Lafon
Jan 4 at 23:05
I think there is a small mistake: We have $I_n = (J_{n+1} color{red}{+} J_{n-1})/2$, which turns the final result into $I_n (x) = pi cos(x) frac{sin(n x)}{sin(x)}$ .
– ComplexYetTrivial
Jan 5 at 8:27
I think there is a small mistake: We have $I_n = (J_{n+1} color{red}{+} J_{n-1})/2$, which turns the final result into $I_n (x) = pi cos(x) frac{sin(n x)}{sin(x)}$ .
– ComplexYetTrivial
Jan 5 at 8:27
Thanks for catching this! I updated the answer.
– Stefan Lafon
Jan 6 at 0:40
Thanks for catching this! I updated the answer.
– Stefan Lafon
Jan 6 at 0:40
add a comment |
A nice way to evaluate generalized integrals is to consider them as the coefficients of an infinite series. Therefore, the coefficient of the $n$th term is simply the integral under question. Before we begin though, there is one identity to note:
$$sumlimits_{ngeq0}z^ncos nx=frac {1-zcos x}{z^2-2zcos x+1}$$
Proof: Rewrite $cos nx$ as the real part of $e^{nix}$. Using the infinite geometric sequences, we get that$$sumlimits_{ngeq0}left(ze^{ix}right)^n=frac 1{1-ze^{ix}}$$Now, take the real part of both sides. Clearly, the left - hand side becomes $z^ncos nx$. Meanwhile, the right - hand side becomes, through some clever rationalization$$begin{align*}operatorname{Re}left[frac 1{1-ze^{ix}}right] & =operatorname{Re}left[frac 1{1-zcos x-zisin x}right]\ & =operatorname{Re}left[frac {1-zcos x+zisin x}{(1-zcos x)^2+z^2sin^2x}right]\ & =frac {1-zcos x}{z^2-2zcos x+1}end{align*}$$completing the proof.
With that in mind, we are ready to begin. Since the OP has stated in the comments that he is trying to evaluate the integral$$I_n=intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$I will show a way to evaluate it in this answer. To wit, denote the generating function of the integral as $G(z)$
$$G(z)=sumlimits_{ngeq0}I_nz^n$$
And remember that the coefficient of $z^n$ simply gives $I_n$. Interchange the sum and the integral, and using the identity we've derived above, get
$$begin{align*}G(z) & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}sumlimits_{ngeq0}z^nbiggr[cos nx-cos ntbiggr]\ & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}left[frac {1-zcos x}{z^2-2zcos x+1}-frac {1-zcos t}{z^2-2zcos t+1}right]end{align*}$$
Combining the two fractions, and recalling that any terms in $z$ are constants, the function becomes
$$G(z)=frac {z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{pi}frac {mathrm dt}{z^2-2zcos t+1}$$
The remaining integral can be easily evaluated using a Weierstrass substitution. Substitute $w=tanleft(tfrac t2right)$ so that
$$begin{array}{|c|c|c|}hline w=tanleft(dfrac t2right) & mathrm dt=dfrac {2,mathrm dw}{1+w^2} & cos t=dfrac {1-w^2}{1+w^2}\hlineend{array}$$
The remaining rational function can be evaluated in an elementary fashion
$$begin{align*}G(z) & =frac {2z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{infty}frac {mathrm dw}{w^2(1+z)^2+(1-z)^2}\ & =frac {2z}{z^2-2zcos x+1}arctanleft(frac {1+z}{1-z}wright),Biggrrvert_0^{infty}\ & =frac {pi z}{z^2-2zcos x+1}end{align*}$$
From the second line, it's important to observe that the argument of the inverse tangent will remain positive if and only if $|z|<1$. When $|z|<1$, then the denominator is positive, as well as the numerator. Therefore, $tfrac {1+z}{1-z}>0$. However, if $|z|>1$, then the argument is less than zero and there is an extra negative sign. For the purpose of this question, we'll consider when $|z|<1$.
Now all we have to do is find the coefficient of $z^n$. There is a nice and convenient way to do this by using
$$2cos x=e^{ix}+e^{-ix}$$
Factoring the denominator by grouping gives
$$begin{align*}frac z{z^2-2zcos x+1} & =frac z{(1-ze^{ix})(1-ze^{-ix})}\ & =zsumlimits_{kgeq0}z^k e^{kix}sumlimits_{lgeq0}z^l e^{-lix}end{align*}$$
Now observe what happens when we expand the products together$$begin{multline}(1+ze^{ix}+z^2e^{2ix}+cdots)(1+ze^{-ix}+z^2e^{-ix}+cdots)\=1+z(e^{ix}+e^{-ix})+z^2(e^{2ix}+1+e^{-2ix})+cdotsend{multline}$$
The sum within the parenthesis seems to start off at the index of the $n$th term and decrease by a factor of two! Using this, it's possible to rewrite the coefficients conveniently as
$$a_k=sumlimits_{m=0}^ke^{(k-2m)ix}=frac {sin x(k+1)}{sin x}$$
Hence$$frac {pi z}{z^2-2zcos x+1}=pisumlimits_{kgeq1}frac {sin xk}{sin x}z^k$$
And setting $k$ as $n$ gives the term $z^n$. Therefore, our integral is simply$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}color{blue}{=frac {pisin xn}{sin x}}$$
@StefanLafon I'm having a little bit of trouble with that; more than I thought initially. I'm trying to mold the infinite sum into$$frac 1{z^2-2zcos x+1}$$on the right - hand side and a $z^n$ term on the left. It's not going very well I must say
– Frank W.
Jan 4 at 18:00
Impressive. To find the coefficients, notice that $frac {2isin x} {z^2-2xcos(x)+1} = frac 1 {z+e^{-ix}}-frac 1 {z+e^{ix}}$.
– Stefan Lafon
Jan 4 at 18:01
Correct the sign mistake in previous comment and show the $n$th integral is $$I_n=pifrac{sin nx}{sin x}$$
– Did
Jan 4 at 18:05
Thanks a lot for your time and effort. This is a very interesting solution but the use of generating functions seems a bit overkill and I don't understand all the details, as I have not yet studied generating functions in details. This problem is supposed to be a medium difficulty problem of first year at university . The tip below the problem suggested to calculate the integral in the question.
– aleph0
Jan 4 at 18:10
1
@Zacky Oh okay. I missed it.
– Frank W.
Jan 4 at 18:55
|
show 3 more comments
A nice way to evaluate generalized integrals is to consider them as the coefficients of an infinite series. Therefore, the coefficient of the $n$th term is simply the integral under question. Before we begin though, there is one identity to note:
$$sumlimits_{ngeq0}z^ncos nx=frac {1-zcos x}{z^2-2zcos x+1}$$
Proof: Rewrite $cos nx$ as the real part of $e^{nix}$. Using the infinite geometric sequences, we get that$$sumlimits_{ngeq0}left(ze^{ix}right)^n=frac 1{1-ze^{ix}}$$Now, take the real part of both sides. Clearly, the left - hand side becomes $z^ncos nx$. Meanwhile, the right - hand side becomes, through some clever rationalization$$begin{align*}operatorname{Re}left[frac 1{1-ze^{ix}}right] & =operatorname{Re}left[frac 1{1-zcos x-zisin x}right]\ & =operatorname{Re}left[frac {1-zcos x+zisin x}{(1-zcos x)^2+z^2sin^2x}right]\ & =frac {1-zcos x}{z^2-2zcos x+1}end{align*}$$completing the proof.
With that in mind, we are ready to begin. Since the OP has stated in the comments that he is trying to evaluate the integral$$I_n=intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$I will show a way to evaluate it in this answer. To wit, denote the generating function of the integral as $G(z)$
$$G(z)=sumlimits_{ngeq0}I_nz^n$$
And remember that the coefficient of $z^n$ simply gives $I_n$. Interchange the sum and the integral, and using the identity we've derived above, get
$$begin{align*}G(z) & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}sumlimits_{ngeq0}z^nbiggr[cos nx-cos ntbiggr]\ & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}left[frac {1-zcos x}{z^2-2zcos x+1}-frac {1-zcos t}{z^2-2zcos t+1}right]end{align*}$$
Combining the two fractions, and recalling that any terms in $z$ are constants, the function becomes
$$G(z)=frac {z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{pi}frac {mathrm dt}{z^2-2zcos t+1}$$
The remaining integral can be easily evaluated using a Weierstrass substitution. Substitute $w=tanleft(tfrac t2right)$ so that
$$begin{array}{|c|c|c|}hline w=tanleft(dfrac t2right) & mathrm dt=dfrac {2,mathrm dw}{1+w^2} & cos t=dfrac {1-w^2}{1+w^2}\hlineend{array}$$
The remaining rational function can be evaluated in an elementary fashion
$$begin{align*}G(z) & =frac {2z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{infty}frac {mathrm dw}{w^2(1+z)^2+(1-z)^2}\ & =frac {2z}{z^2-2zcos x+1}arctanleft(frac {1+z}{1-z}wright),Biggrrvert_0^{infty}\ & =frac {pi z}{z^2-2zcos x+1}end{align*}$$
From the second line, it's important to observe that the argument of the inverse tangent will remain positive if and only if $|z|<1$. When $|z|<1$, then the denominator is positive, as well as the numerator. Therefore, $tfrac {1+z}{1-z}>0$. However, if $|z|>1$, then the argument is less than zero and there is an extra negative sign. For the purpose of this question, we'll consider when $|z|<1$.
Now all we have to do is find the coefficient of $z^n$. There is a nice and convenient way to do this by using
$$2cos x=e^{ix}+e^{-ix}$$
Factoring the denominator by grouping gives
$$begin{align*}frac z{z^2-2zcos x+1} & =frac z{(1-ze^{ix})(1-ze^{-ix})}\ & =zsumlimits_{kgeq0}z^k e^{kix}sumlimits_{lgeq0}z^l e^{-lix}end{align*}$$
Now observe what happens when we expand the products together$$begin{multline}(1+ze^{ix}+z^2e^{2ix}+cdots)(1+ze^{-ix}+z^2e^{-ix}+cdots)\=1+z(e^{ix}+e^{-ix})+z^2(e^{2ix}+1+e^{-2ix})+cdotsend{multline}$$
The sum within the parenthesis seems to start off at the index of the $n$th term and decrease by a factor of two! Using this, it's possible to rewrite the coefficients conveniently as
$$a_k=sumlimits_{m=0}^ke^{(k-2m)ix}=frac {sin x(k+1)}{sin x}$$
Hence$$frac {pi z}{z^2-2zcos x+1}=pisumlimits_{kgeq1}frac {sin xk}{sin x}z^k$$
And setting $k$ as $n$ gives the term $z^n$. Therefore, our integral is simply$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}color{blue}{=frac {pisin xn}{sin x}}$$
@StefanLafon I'm having a little bit of trouble with that; more than I thought initially. I'm trying to mold the infinite sum into$$frac 1{z^2-2zcos x+1}$$on the right - hand side and a $z^n$ term on the left. It's not going very well I must say
– Frank W.
Jan 4 at 18:00
Impressive. To find the coefficients, notice that $frac {2isin x} {z^2-2xcos(x)+1} = frac 1 {z+e^{-ix}}-frac 1 {z+e^{ix}}$.
– Stefan Lafon
Jan 4 at 18:01
Correct the sign mistake in previous comment and show the $n$th integral is $$I_n=pifrac{sin nx}{sin x}$$
– Did
Jan 4 at 18:05
Thanks a lot for your time and effort. This is a very interesting solution but the use of generating functions seems a bit overkill and I don't understand all the details, as I have not yet studied generating functions in details. This problem is supposed to be a medium difficulty problem of first year at university . The tip below the problem suggested to calculate the integral in the question.
– aleph0
Jan 4 at 18:10
1
@Zacky Oh okay. I missed it.
– Frank W.
Jan 4 at 18:55
|
show 3 more comments
A nice way to evaluate generalized integrals is to consider them as the coefficients of an infinite series. Therefore, the coefficient of the $n$th term is simply the integral under question. Before we begin though, there is one identity to note:
$$sumlimits_{ngeq0}z^ncos nx=frac {1-zcos x}{z^2-2zcos x+1}$$
Proof: Rewrite $cos nx$ as the real part of $e^{nix}$. Using the infinite geometric sequences, we get that$$sumlimits_{ngeq0}left(ze^{ix}right)^n=frac 1{1-ze^{ix}}$$Now, take the real part of both sides. Clearly, the left - hand side becomes $z^ncos nx$. Meanwhile, the right - hand side becomes, through some clever rationalization$$begin{align*}operatorname{Re}left[frac 1{1-ze^{ix}}right] & =operatorname{Re}left[frac 1{1-zcos x-zisin x}right]\ & =operatorname{Re}left[frac {1-zcos x+zisin x}{(1-zcos x)^2+z^2sin^2x}right]\ & =frac {1-zcos x}{z^2-2zcos x+1}end{align*}$$completing the proof.
With that in mind, we are ready to begin. Since the OP has stated in the comments that he is trying to evaluate the integral$$I_n=intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$I will show a way to evaluate it in this answer. To wit, denote the generating function of the integral as $G(z)$
$$G(z)=sumlimits_{ngeq0}I_nz^n$$
And remember that the coefficient of $z^n$ simply gives $I_n$. Interchange the sum and the integral, and using the identity we've derived above, get
$$begin{align*}G(z) & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}sumlimits_{ngeq0}z^nbiggr[cos nx-cos ntbiggr]\ & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}left[frac {1-zcos x}{z^2-2zcos x+1}-frac {1-zcos t}{z^2-2zcos t+1}right]end{align*}$$
Combining the two fractions, and recalling that any terms in $z$ are constants, the function becomes
$$G(z)=frac {z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{pi}frac {mathrm dt}{z^2-2zcos t+1}$$
The remaining integral can be easily evaluated using a Weierstrass substitution. Substitute $w=tanleft(tfrac t2right)$ so that
$$begin{array}{|c|c|c|}hline w=tanleft(dfrac t2right) & mathrm dt=dfrac {2,mathrm dw}{1+w^2} & cos t=dfrac {1-w^2}{1+w^2}\hlineend{array}$$
The remaining rational function can be evaluated in an elementary fashion
$$begin{align*}G(z) & =frac {2z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{infty}frac {mathrm dw}{w^2(1+z)^2+(1-z)^2}\ & =frac {2z}{z^2-2zcos x+1}arctanleft(frac {1+z}{1-z}wright),Biggrrvert_0^{infty}\ & =frac {pi z}{z^2-2zcos x+1}end{align*}$$
From the second line, it's important to observe that the argument of the inverse tangent will remain positive if and only if $|z|<1$. When $|z|<1$, then the denominator is positive, as well as the numerator. Therefore, $tfrac {1+z}{1-z}>0$. However, if $|z|>1$, then the argument is less than zero and there is an extra negative sign. For the purpose of this question, we'll consider when $|z|<1$.
Now all we have to do is find the coefficient of $z^n$. There is a nice and convenient way to do this by using
$$2cos x=e^{ix}+e^{-ix}$$
Factoring the denominator by grouping gives
$$begin{align*}frac z{z^2-2zcos x+1} & =frac z{(1-ze^{ix})(1-ze^{-ix})}\ & =zsumlimits_{kgeq0}z^k e^{kix}sumlimits_{lgeq0}z^l e^{-lix}end{align*}$$
Now observe what happens when we expand the products together$$begin{multline}(1+ze^{ix}+z^2e^{2ix}+cdots)(1+ze^{-ix}+z^2e^{-ix}+cdots)\=1+z(e^{ix}+e^{-ix})+z^2(e^{2ix}+1+e^{-2ix})+cdotsend{multline}$$
The sum within the parenthesis seems to start off at the index of the $n$th term and decrease by a factor of two! Using this, it's possible to rewrite the coefficients conveniently as
$$a_k=sumlimits_{m=0}^ke^{(k-2m)ix}=frac {sin x(k+1)}{sin x}$$
Hence$$frac {pi z}{z^2-2zcos x+1}=pisumlimits_{kgeq1}frac {sin xk}{sin x}z^k$$
And setting $k$ as $n$ gives the term $z^n$. Therefore, our integral is simply$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}color{blue}{=frac {pisin xn}{sin x}}$$
A nice way to evaluate generalized integrals is to consider them as the coefficients of an infinite series. Therefore, the coefficient of the $n$th term is simply the integral under question. Before we begin though, there is one identity to note:
$$sumlimits_{ngeq0}z^ncos nx=frac {1-zcos x}{z^2-2zcos x+1}$$
Proof: Rewrite $cos nx$ as the real part of $e^{nix}$. Using the infinite geometric sequences, we get that$$sumlimits_{ngeq0}left(ze^{ix}right)^n=frac 1{1-ze^{ix}}$$Now, take the real part of both sides. Clearly, the left - hand side becomes $z^ncos nx$. Meanwhile, the right - hand side becomes, through some clever rationalization$$begin{align*}operatorname{Re}left[frac 1{1-ze^{ix}}right] & =operatorname{Re}left[frac 1{1-zcos x-zisin x}right]\ & =operatorname{Re}left[frac {1-zcos x+zisin x}{(1-zcos x)^2+z^2sin^2x}right]\ & =frac {1-zcos x}{z^2-2zcos x+1}end{align*}$$completing the proof.
With that in mind, we are ready to begin. Since the OP has stated in the comments that he is trying to evaluate the integral$$I_n=intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$I will show a way to evaluate it in this answer. To wit, denote the generating function of the integral as $G(z)$
$$G(z)=sumlimits_{ngeq0}I_nz^n$$
And remember that the coefficient of $z^n$ simply gives $I_n$. Interchange the sum and the integral, and using the identity we've derived above, get
$$begin{align*}G(z) & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}sumlimits_{ngeq0}z^nbiggr[cos nx-cos ntbiggr]\ & =intlimits_0^{pi}frac {mathrm dt}{cos x-cos t}left[frac {1-zcos x}{z^2-2zcos x+1}-frac {1-zcos t}{z^2-2zcos t+1}right]end{align*}$$
Combining the two fractions, and recalling that any terms in $z$ are constants, the function becomes
$$G(z)=frac {z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{pi}frac {mathrm dt}{z^2-2zcos t+1}$$
The remaining integral can be easily evaluated using a Weierstrass substitution. Substitute $w=tanleft(tfrac t2right)$ so that
$$begin{array}{|c|c|c|}hline w=tanleft(dfrac t2right) & mathrm dt=dfrac {2,mathrm dw}{1+w^2} & cos t=dfrac {1-w^2}{1+w^2}\hlineend{array}$$
The remaining rational function can be evaluated in an elementary fashion
$$begin{align*}G(z) & =frac {2z(1-z^2)}{z^2-2zcos x+1}intlimits_0^{infty}frac {mathrm dw}{w^2(1+z)^2+(1-z)^2}\ & =frac {2z}{z^2-2zcos x+1}arctanleft(frac {1+z}{1-z}wright),Biggrrvert_0^{infty}\ & =frac {pi z}{z^2-2zcos x+1}end{align*}$$
From the second line, it's important to observe that the argument of the inverse tangent will remain positive if and only if $|z|<1$. When $|z|<1$, then the denominator is positive, as well as the numerator. Therefore, $tfrac {1+z}{1-z}>0$. However, if $|z|>1$, then the argument is less than zero and there is an extra negative sign. For the purpose of this question, we'll consider when $|z|<1$.
Now all we have to do is find the coefficient of $z^n$. There is a nice and convenient way to do this by using
$$2cos x=e^{ix}+e^{-ix}$$
Factoring the denominator by grouping gives
$$begin{align*}frac z{z^2-2zcos x+1} & =frac z{(1-ze^{ix})(1-ze^{-ix})}\ & =zsumlimits_{kgeq0}z^k e^{kix}sumlimits_{lgeq0}z^l e^{-lix}end{align*}$$
Now observe what happens when we expand the products together$$begin{multline}(1+ze^{ix}+z^2e^{2ix}+cdots)(1+ze^{-ix}+z^2e^{-ix}+cdots)\=1+z(e^{ix}+e^{-ix})+z^2(e^{2ix}+1+e^{-2ix})+cdotsend{multline}$$
The sum within the parenthesis seems to start off at the index of the $n$th term and decrease by a factor of two! Using this, it's possible to rewrite the coefficients conveniently as
$$a_k=sumlimits_{m=0}^ke^{(k-2m)ix}=frac {sin x(k+1)}{sin x}$$
Hence$$frac {pi z}{z^2-2zcos x+1}=pisumlimits_{kgeq1}frac {sin xk}{sin x}z^k$$
And setting $k$ as $n$ gives the term $z^n$. Therefore, our integral is simply$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}color{blue}{=frac {pisin xn}{sin x}}$$
edited Jan 4 at 22:46
answered Jan 4 at 17:54
Frank W.Frank W.
3,1891321
3,1891321
@StefanLafon I'm having a little bit of trouble with that; more than I thought initially. I'm trying to mold the infinite sum into$$frac 1{z^2-2zcos x+1}$$on the right - hand side and a $z^n$ term on the left. It's not going very well I must say
– Frank W.
Jan 4 at 18:00
Impressive. To find the coefficients, notice that $frac {2isin x} {z^2-2xcos(x)+1} = frac 1 {z+e^{-ix}}-frac 1 {z+e^{ix}}$.
– Stefan Lafon
Jan 4 at 18:01
Correct the sign mistake in previous comment and show the $n$th integral is $$I_n=pifrac{sin nx}{sin x}$$
– Did
Jan 4 at 18:05
Thanks a lot for your time and effort. This is a very interesting solution but the use of generating functions seems a bit overkill and I don't understand all the details, as I have not yet studied generating functions in details. This problem is supposed to be a medium difficulty problem of first year at university . The tip below the problem suggested to calculate the integral in the question.
– aleph0
Jan 4 at 18:10
1
@Zacky Oh okay. I missed it.
– Frank W.
Jan 4 at 18:55
|
show 3 more comments
@StefanLafon I'm having a little bit of trouble with that; more than I thought initially. I'm trying to mold the infinite sum into$$frac 1{z^2-2zcos x+1}$$on the right - hand side and a $z^n$ term on the left. It's not going very well I must say
– Frank W.
Jan 4 at 18:00
Impressive. To find the coefficients, notice that $frac {2isin x} {z^2-2xcos(x)+1} = frac 1 {z+e^{-ix}}-frac 1 {z+e^{ix}}$.
– Stefan Lafon
Jan 4 at 18:01
Correct the sign mistake in previous comment and show the $n$th integral is $$I_n=pifrac{sin nx}{sin x}$$
– Did
Jan 4 at 18:05
Thanks a lot for your time and effort. This is a very interesting solution but the use of generating functions seems a bit overkill and I don't understand all the details, as I have not yet studied generating functions in details. This problem is supposed to be a medium difficulty problem of first year at university . The tip below the problem suggested to calculate the integral in the question.
– aleph0
Jan 4 at 18:10
1
@Zacky Oh okay. I missed it.
– Frank W.
Jan 4 at 18:55
@StefanLafon I'm having a little bit of trouble with that; more than I thought initially. I'm trying to mold the infinite sum into$$frac 1{z^2-2zcos x+1}$$on the right - hand side and a $z^n$ term on the left. It's not going very well I must say
– Frank W.
Jan 4 at 18:00
@StefanLafon I'm having a little bit of trouble with that; more than I thought initially. I'm trying to mold the infinite sum into$$frac 1{z^2-2zcos x+1}$$on the right - hand side and a $z^n$ term on the left. It's not going very well I must say
– Frank W.
Jan 4 at 18:00
Impressive. To find the coefficients, notice that $frac {2isin x} {z^2-2xcos(x)+1} = frac 1 {z+e^{-ix}}-frac 1 {z+e^{ix}}$.
– Stefan Lafon
Jan 4 at 18:01
Impressive. To find the coefficients, notice that $frac {2isin x} {z^2-2xcos(x)+1} = frac 1 {z+e^{-ix}}-frac 1 {z+e^{ix}}$.
– Stefan Lafon
Jan 4 at 18:01
Correct the sign mistake in previous comment and show the $n$th integral is $$I_n=pifrac{sin nx}{sin x}$$
– Did
Jan 4 at 18:05
Correct the sign mistake in previous comment and show the $n$th integral is $$I_n=pifrac{sin nx}{sin x}$$
– Did
Jan 4 at 18:05
Thanks a lot for your time and effort. This is a very interesting solution but the use of generating functions seems a bit overkill and I don't understand all the details, as I have not yet studied generating functions in details. This problem is supposed to be a medium difficulty problem of first year at university . The tip below the problem suggested to calculate the integral in the question.
– aleph0
Jan 4 at 18:10
Thanks a lot for your time and effort. This is a very interesting solution but the use of generating functions seems a bit overkill and I don't understand all the details, as I have not yet studied generating functions in details. This problem is supposed to be a medium difficulty problem of first year at university . The tip below the problem suggested to calculate the integral in the question.
– aleph0
Jan 4 at 18:10
1
1
@Zacky Oh okay. I missed it.
– Frank W.
Jan 4 at 18:55
@Zacky Oh okay. I missed it.
– Frank W.
Jan 4 at 18:55
|
show 3 more comments
Completing Frank's solution:
$$ [z^n]frac{pi z}{z^2-2zcos x+1} = frac{pi}{2}[z^{n}]left(frac{1}{z-e^{ix}}+frac{1}{z-e^{-ix}}right) $$
equals, by geometric series,
$$ frac{pi}{2}left(-e^{-(n+1)ix}-e^{(n+1)ix}right)=-picos((n+1)x). $$
add a comment |
Completing Frank's solution:
$$ [z^n]frac{pi z}{z^2-2zcos x+1} = frac{pi}{2}[z^{n}]left(frac{1}{z-e^{ix}}+frac{1}{z-e^{-ix}}right) $$
equals, by geometric series,
$$ frac{pi}{2}left(-e^{-(n+1)ix}-e^{(n+1)ix}right)=-picos((n+1)x). $$
add a comment |
Completing Frank's solution:
$$ [z^n]frac{pi z}{z^2-2zcos x+1} = frac{pi}{2}[z^{n}]left(frac{1}{z-e^{ix}}+frac{1}{z-e^{-ix}}right) $$
equals, by geometric series,
$$ frac{pi}{2}left(-e^{-(n+1)ix}-e^{(n+1)ix}right)=-picos((n+1)x). $$
Completing Frank's solution:
$$ [z^n]frac{pi z}{z^2-2zcos x+1} = frac{pi}{2}[z^{n}]left(frac{1}{z-e^{ix}}+frac{1}{z-e^{-ix}}right) $$
equals, by geometric series,
$$ frac{pi}{2}left(-e^{-(n+1)ix}-e^{(n+1)ix}right)=-picos((n+1)x). $$
answered Jan 4 at 18:20
Jack D'AurizioJack D'Aurizio
287k33280659
287k33280659
add a comment |
add a comment |
I will do the integral for $n=3$, unfortunately this will disprove your given result. (unless I did a mistake)
$$I_3= int_0^pi frac{cos(3x)cos x -cos(3t)cos t}{cos x-cos t}dt$$
Since $cos(3 y)=4cos^3 y -3cos y$ we have: $$cos(3x)cos x -cos (3t)cos t=4(cos^4 x-cos^4t)-3(cos^2 x-cos^2 t)$$
$$=4(cos x-cos t)(cos x+cos t) (cos^2 x+cos^2t)-3(cos x-cos t)(cos x+cos t)$$
$$Rightarrow I_3=int_0^pi (cos x+cos t)(4(cos ^2 x+cos^2 t)-3)dt$$
$$overset{pi-tto t}=int_0^pi (cos x-cos t)(4(cos^2 x+cos^2 t)-3)dt$$
$$Rightarrow 2I_3=2cos xint_0^pi (4(cos^2 x+cos^2 t)-3)$$
$$Rightarrow I_3= 4pi cos^3 x +cos x underbrace{int_0^pi cos^2 tdt}_{=frac{pi}{2}}-3pi cos x=2pi(cos x+2cos(3x))$$
For $n=4$ we have: $$cos(4x)=8cos^4 x-8cos^2 x+1$$
Denoting $cx=cos x$ and $ct=cos t,$ we get the integrand to be:
$$frac{8(cx^5-ct^5)-8(cx^3-ct^3)+(cx-ct)}{cx-ct}$$
And using the fact that: $$a^5-b^5=(a-b)(a^4+a^3b+a^2b^2+ab^3+b^4$$
$$Rightarrow I_4=8int_0^pi(cx^4+cx^3 ct+cx^2 ct^2 +cx ct^3+ct^4)dt$$
$$-8int_0^pi (cx^2+cxct+ct^2) dt +int_0^pi dt$$
We have: $$int_0^pi ct dt=int_0^pi ct^3 dt =0$$
$$int_0^pi ct^2 dt= frac{pi}{2}, int_0^pi ct^4 dt=frac{3pi}{8}$$
$$Rightarrow I_4=(8pi cx^4 +4pi cx^2 +3pi )-(8pi cx^2 +4pi)+pi$$
$$=8pi cx^4 -4pi cx^2 =4pi cos^2 x cos(2x)$$
1
I'm getting a slightly different answer too than what the OP conjectured.
– Frank W.
Jan 4 at 18:00
1
It seems like it only works for $n=1$ and $n=2$ until now. Note that even $n=0$ fails.
– Zacky
Jan 4 at 18:01
Thanks for doing that. I guess my conjecture was wrong...
– aleph0
Jan 4 at 18:08
add a comment |
I will do the integral for $n=3$, unfortunately this will disprove your given result. (unless I did a mistake)
$$I_3= int_0^pi frac{cos(3x)cos x -cos(3t)cos t}{cos x-cos t}dt$$
Since $cos(3 y)=4cos^3 y -3cos y$ we have: $$cos(3x)cos x -cos (3t)cos t=4(cos^4 x-cos^4t)-3(cos^2 x-cos^2 t)$$
$$=4(cos x-cos t)(cos x+cos t) (cos^2 x+cos^2t)-3(cos x-cos t)(cos x+cos t)$$
$$Rightarrow I_3=int_0^pi (cos x+cos t)(4(cos ^2 x+cos^2 t)-3)dt$$
$$overset{pi-tto t}=int_0^pi (cos x-cos t)(4(cos^2 x+cos^2 t)-3)dt$$
$$Rightarrow 2I_3=2cos xint_0^pi (4(cos^2 x+cos^2 t)-3)$$
$$Rightarrow I_3= 4pi cos^3 x +cos x underbrace{int_0^pi cos^2 tdt}_{=frac{pi}{2}}-3pi cos x=2pi(cos x+2cos(3x))$$
For $n=4$ we have: $$cos(4x)=8cos^4 x-8cos^2 x+1$$
Denoting $cx=cos x$ and $ct=cos t,$ we get the integrand to be:
$$frac{8(cx^5-ct^5)-8(cx^3-ct^3)+(cx-ct)}{cx-ct}$$
And using the fact that: $$a^5-b^5=(a-b)(a^4+a^3b+a^2b^2+ab^3+b^4$$
$$Rightarrow I_4=8int_0^pi(cx^4+cx^3 ct+cx^2 ct^2 +cx ct^3+ct^4)dt$$
$$-8int_0^pi (cx^2+cxct+ct^2) dt +int_0^pi dt$$
We have: $$int_0^pi ct dt=int_0^pi ct^3 dt =0$$
$$int_0^pi ct^2 dt= frac{pi}{2}, int_0^pi ct^4 dt=frac{3pi}{8}$$
$$Rightarrow I_4=(8pi cx^4 +4pi cx^2 +3pi )-(8pi cx^2 +4pi)+pi$$
$$=8pi cx^4 -4pi cx^2 =4pi cos^2 x cos(2x)$$
1
I'm getting a slightly different answer too than what the OP conjectured.
– Frank W.
Jan 4 at 18:00
1
It seems like it only works for $n=1$ and $n=2$ until now. Note that even $n=0$ fails.
– Zacky
Jan 4 at 18:01
Thanks for doing that. I guess my conjecture was wrong...
– aleph0
Jan 4 at 18:08
add a comment |
I will do the integral for $n=3$, unfortunately this will disprove your given result. (unless I did a mistake)
$$I_3= int_0^pi frac{cos(3x)cos x -cos(3t)cos t}{cos x-cos t}dt$$
Since $cos(3 y)=4cos^3 y -3cos y$ we have: $$cos(3x)cos x -cos (3t)cos t=4(cos^4 x-cos^4t)-3(cos^2 x-cos^2 t)$$
$$=4(cos x-cos t)(cos x+cos t) (cos^2 x+cos^2t)-3(cos x-cos t)(cos x+cos t)$$
$$Rightarrow I_3=int_0^pi (cos x+cos t)(4(cos ^2 x+cos^2 t)-3)dt$$
$$overset{pi-tto t}=int_0^pi (cos x-cos t)(4(cos^2 x+cos^2 t)-3)dt$$
$$Rightarrow 2I_3=2cos xint_0^pi (4(cos^2 x+cos^2 t)-3)$$
$$Rightarrow I_3= 4pi cos^3 x +cos x underbrace{int_0^pi cos^2 tdt}_{=frac{pi}{2}}-3pi cos x=2pi(cos x+2cos(3x))$$
For $n=4$ we have: $$cos(4x)=8cos^4 x-8cos^2 x+1$$
Denoting $cx=cos x$ and $ct=cos t,$ we get the integrand to be:
$$frac{8(cx^5-ct^5)-8(cx^3-ct^3)+(cx-ct)}{cx-ct}$$
And using the fact that: $$a^5-b^5=(a-b)(a^4+a^3b+a^2b^2+ab^3+b^4$$
$$Rightarrow I_4=8int_0^pi(cx^4+cx^3 ct+cx^2 ct^2 +cx ct^3+ct^4)dt$$
$$-8int_0^pi (cx^2+cxct+ct^2) dt +int_0^pi dt$$
We have: $$int_0^pi ct dt=int_0^pi ct^3 dt =0$$
$$int_0^pi ct^2 dt= frac{pi}{2}, int_0^pi ct^4 dt=frac{3pi}{8}$$
$$Rightarrow I_4=(8pi cx^4 +4pi cx^2 +3pi )-(8pi cx^2 +4pi)+pi$$
$$=8pi cx^4 -4pi cx^2 =4pi cos^2 x cos(2x)$$
I will do the integral for $n=3$, unfortunately this will disprove your given result. (unless I did a mistake)
$$I_3= int_0^pi frac{cos(3x)cos x -cos(3t)cos t}{cos x-cos t}dt$$
Since $cos(3 y)=4cos^3 y -3cos y$ we have: $$cos(3x)cos x -cos (3t)cos t=4(cos^4 x-cos^4t)-3(cos^2 x-cos^2 t)$$
$$=4(cos x-cos t)(cos x+cos t) (cos^2 x+cos^2t)-3(cos x-cos t)(cos x+cos t)$$
$$Rightarrow I_3=int_0^pi (cos x+cos t)(4(cos ^2 x+cos^2 t)-3)dt$$
$$overset{pi-tto t}=int_0^pi (cos x-cos t)(4(cos^2 x+cos^2 t)-3)dt$$
$$Rightarrow 2I_3=2cos xint_0^pi (4(cos^2 x+cos^2 t)-3)$$
$$Rightarrow I_3= 4pi cos^3 x +cos x underbrace{int_0^pi cos^2 tdt}_{=frac{pi}{2}}-3pi cos x=2pi(cos x+2cos(3x))$$
For $n=4$ we have: $$cos(4x)=8cos^4 x-8cos^2 x+1$$
Denoting $cx=cos x$ and $ct=cos t,$ we get the integrand to be:
$$frac{8(cx^5-ct^5)-8(cx^3-ct^3)+(cx-ct)}{cx-ct}$$
And using the fact that: $$a^5-b^5=(a-b)(a^4+a^3b+a^2b^2+ab^3+b^4$$
$$Rightarrow I_4=8int_0^pi(cx^4+cx^3 ct+cx^2 ct^2 +cx ct^3+ct^4)dt$$
$$-8int_0^pi (cx^2+cxct+ct^2) dt +int_0^pi dt$$
We have: $$int_0^pi ct dt=int_0^pi ct^3 dt =0$$
$$int_0^pi ct^2 dt= frac{pi}{2}, int_0^pi ct^4 dt=frac{3pi}{8}$$
$$Rightarrow I_4=(8pi cx^4 +4pi cx^2 +3pi )-(8pi cx^2 +4pi)+pi$$
$$=8pi cx^4 -4pi cx^2 =4pi cos^2 x cos(2x)$$
edited Jan 4 at 18:29
answered Jan 4 at 17:56
ZackyZacky
5,1381752
5,1381752
1
I'm getting a slightly different answer too than what the OP conjectured.
– Frank W.
Jan 4 at 18:00
1
It seems like it only works for $n=1$ and $n=2$ until now. Note that even $n=0$ fails.
– Zacky
Jan 4 at 18:01
Thanks for doing that. I guess my conjecture was wrong...
– aleph0
Jan 4 at 18:08
add a comment |
1
I'm getting a slightly different answer too than what the OP conjectured.
– Frank W.
Jan 4 at 18:00
1
It seems like it only works for $n=1$ and $n=2$ until now. Note that even $n=0$ fails.
– Zacky
Jan 4 at 18:01
Thanks for doing that. I guess my conjecture was wrong...
– aleph0
Jan 4 at 18:08
1
1
I'm getting a slightly different answer too than what the OP conjectured.
– Frank W.
Jan 4 at 18:00
I'm getting a slightly different answer too than what the OP conjectured.
– Frank W.
Jan 4 at 18:00
1
1
It seems like it only works for $n=1$ and $n=2$ until now. Note that even $n=0$ fails.
– Zacky
Jan 4 at 18:01
It seems like it only works for $n=1$ and $n=2$ until now. Note that even $n=0$ fails.
– Zacky
Jan 4 at 18:01
Thanks for doing that. I guess my conjecture was wrong...
– aleph0
Jan 4 at 18:08
Thanks for doing that. I guess my conjecture was wrong...
– aleph0
Jan 4 at 18:08
add a comment |
An alternative solution to the problem:
For $n in mathbb{N}$ and $x in (0,pi)$ define
$$J_n (x) equiv int limits_0^pi frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t , . $$
We can use the identities ($(2)$ follows from the geometric progression formula)
begin{align}
cos(xi) - cos(tau) &= - 2 sin left(frac{xi + tau}{2}right) sin left(frac{xi - tau}{2}right) , , , xi,tau in mathbb{R} , , tag{1} \
frac{sin(n y)}{sin(y)} &= mathrm{e}^{-mathrm{i}(n-1)y} sum limits_{k=0}^{n-1} mathrm{e}^{2mathrm{i} k y} , , , n in mathbb{N} , , , y in mathbb{R} , , tag{2} \
int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t &= 2 pi delta_{k,l} , , , k,l in mathbb{Z} , , tag{3}
end{align}
to compute
begin{align}
J_n (x) &= frac{1}{2} int limits_0^{2pi} frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t stackrel{(1)}{=} frac{1}{2} int limits_0^{2pi} frac{sin left(nfrac{x+t}{2}right)}{sin left(frac{x+t}{2}right)} frac{sin left(nfrac{x-t}{2}right)}{sin left(frac{x-t}{2}right)} , mathrm{d} t \
&stackrel{(2)}{=} frac{1}{2} mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k,l=0}^{n-1} mathrm{e}^{mathrm{i} (k+l) x} int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t stackrel{(3)}{=} pi mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k=0}^{n-1} mathrm{e}^{2 mathrm{i} k x} \
&stackrel{(2)}{=} pi frac{sin(nx)}{sin(x)} , .
end{align}
This result directly leads to
begin{align}
I_n(x) &equiv int limits_0^pi frac{cos(n x) cos(x) - cos(n t) cos(t)}{cos(x) - cos(t)} , mathrm{d} t = int limits_0^pi left[cos(x)frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} + cos(n t)right], mathrm{d} t \
&= cos(x) J_n(x) + 0 = pi cos(x) frac{sin(nx)}{sin(x)} , .
end{align}
Great solution ! Thank you.
– aleph0
Jan 5 at 11:48
add a comment |
An alternative solution to the problem:
For $n in mathbb{N}$ and $x in (0,pi)$ define
$$J_n (x) equiv int limits_0^pi frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t , . $$
We can use the identities ($(2)$ follows from the geometric progression formula)
begin{align}
cos(xi) - cos(tau) &= - 2 sin left(frac{xi + tau}{2}right) sin left(frac{xi - tau}{2}right) , , , xi,tau in mathbb{R} , , tag{1} \
frac{sin(n y)}{sin(y)} &= mathrm{e}^{-mathrm{i}(n-1)y} sum limits_{k=0}^{n-1} mathrm{e}^{2mathrm{i} k y} , , , n in mathbb{N} , , , y in mathbb{R} , , tag{2} \
int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t &= 2 pi delta_{k,l} , , , k,l in mathbb{Z} , , tag{3}
end{align}
to compute
begin{align}
J_n (x) &= frac{1}{2} int limits_0^{2pi} frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t stackrel{(1)}{=} frac{1}{2} int limits_0^{2pi} frac{sin left(nfrac{x+t}{2}right)}{sin left(frac{x+t}{2}right)} frac{sin left(nfrac{x-t}{2}right)}{sin left(frac{x-t}{2}right)} , mathrm{d} t \
&stackrel{(2)}{=} frac{1}{2} mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k,l=0}^{n-1} mathrm{e}^{mathrm{i} (k+l) x} int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t stackrel{(3)}{=} pi mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k=0}^{n-1} mathrm{e}^{2 mathrm{i} k x} \
&stackrel{(2)}{=} pi frac{sin(nx)}{sin(x)} , .
end{align}
This result directly leads to
begin{align}
I_n(x) &equiv int limits_0^pi frac{cos(n x) cos(x) - cos(n t) cos(t)}{cos(x) - cos(t)} , mathrm{d} t = int limits_0^pi left[cos(x)frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} + cos(n t)right], mathrm{d} t \
&= cos(x) J_n(x) + 0 = pi cos(x) frac{sin(nx)}{sin(x)} , .
end{align}
Great solution ! Thank you.
– aleph0
Jan 5 at 11:48
add a comment |
An alternative solution to the problem:
For $n in mathbb{N}$ and $x in (0,pi)$ define
$$J_n (x) equiv int limits_0^pi frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t , . $$
We can use the identities ($(2)$ follows from the geometric progression formula)
begin{align}
cos(xi) - cos(tau) &= - 2 sin left(frac{xi + tau}{2}right) sin left(frac{xi - tau}{2}right) , , , xi,tau in mathbb{R} , , tag{1} \
frac{sin(n y)}{sin(y)} &= mathrm{e}^{-mathrm{i}(n-1)y} sum limits_{k=0}^{n-1} mathrm{e}^{2mathrm{i} k y} , , , n in mathbb{N} , , , y in mathbb{R} , , tag{2} \
int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t &= 2 pi delta_{k,l} , , , k,l in mathbb{Z} , , tag{3}
end{align}
to compute
begin{align}
J_n (x) &= frac{1}{2} int limits_0^{2pi} frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t stackrel{(1)}{=} frac{1}{2} int limits_0^{2pi} frac{sin left(nfrac{x+t}{2}right)}{sin left(frac{x+t}{2}right)} frac{sin left(nfrac{x-t}{2}right)}{sin left(frac{x-t}{2}right)} , mathrm{d} t \
&stackrel{(2)}{=} frac{1}{2} mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k,l=0}^{n-1} mathrm{e}^{mathrm{i} (k+l) x} int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t stackrel{(3)}{=} pi mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k=0}^{n-1} mathrm{e}^{2 mathrm{i} k x} \
&stackrel{(2)}{=} pi frac{sin(nx)}{sin(x)} , .
end{align}
This result directly leads to
begin{align}
I_n(x) &equiv int limits_0^pi frac{cos(n x) cos(x) - cos(n t) cos(t)}{cos(x) - cos(t)} , mathrm{d} t = int limits_0^pi left[cos(x)frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} + cos(n t)right], mathrm{d} t \
&= cos(x) J_n(x) + 0 = pi cos(x) frac{sin(nx)}{sin(x)} , .
end{align}
An alternative solution to the problem:
For $n in mathbb{N}$ and $x in (0,pi)$ define
$$J_n (x) equiv int limits_0^pi frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t , . $$
We can use the identities ($(2)$ follows from the geometric progression formula)
begin{align}
cos(xi) - cos(tau) &= - 2 sin left(frac{xi + tau}{2}right) sin left(frac{xi - tau}{2}right) , , , xi,tau in mathbb{R} , , tag{1} \
frac{sin(n y)}{sin(y)} &= mathrm{e}^{-mathrm{i}(n-1)y} sum limits_{k=0}^{n-1} mathrm{e}^{2mathrm{i} k y} , , , n in mathbb{N} , , , y in mathbb{R} , , tag{2} \
int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t &= 2 pi delta_{k,l} , , , k,l in mathbb{Z} , , tag{3}
end{align}
to compute
begin{align}
J_n (x) &= frac{1}{2} int limits_0^{2pi} frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} , mathrm{d} t stackrel{(1)}{=} frac{1}{2} int limits_0^{2pi} frac{sin left(nfrac{x+t}{2}right)}{sin left(frac{x+t}{2}right)} frac{sin left(nfrac{x-t}{2}right)}{sin left(frac{x-t}{2}right)} , mathrm{d} t \
&stackrel{(2)}{=} frac{1}{2} mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k,l=0}^{n-1} mathrm{e}^{mathrm{i} (k+l) x} int limits_0^{2 pi} mathrm{e}^{mathrm{i}(k-l) t} , mathrm{d} t stackrel{(3)}{=} pi mathrm{e}^{-mathrm{i} (n-1) x} sum limits_{k=0}^{n-1} mathrm{e}^{2 mathrm{i} k x} \
&stackrel{(2)}{=} pi frac{sin(nx)}{sin(x)} , .
end{align}
This result directly leads to
begin{align}
I_n(x) &equiv int limits_0^pi frac{cos(n x) cos(x) - cos(n t) cos(t)}{cos(x) - cos(t)} , mathrm{d} t = int limits_0^pi left[cos(x)frac{cos(n x) - cos(n t)}{cos(x) - cos(t)} + cos(n t)right], mathrm{d} t \
&= cos(x) J_n(x) + 0 = pi cos(x) frac{sin(nx)}{sin(x)} , .
end{align}
answered Jan 5 at 8:55
ComplexYetTrivialComplexYetTrivial
3,5552628
3,5552628
Great solution ! Thank you.
– aleph0
Jan 5 at 11:48
add a comment |
Great solution ! Thank you.
– aleph0
Jan 5 at 11:48
Great solution ! Thank you.
– aleph0
Jan 5 at 11:48
Great solution ! Thank you.
– aleph0
Jan 5 at 11:48
add a comment |
aleph0 is a new contributor. Be nice, and check out our Code of Conduct.
aleph0 is a new contributor. Be nice, and check out our Code of Conduct.
aleph0 is a new contributor. Be nice, and check out our Code of Conduct.
aleph0 is a new contributor. Be nice, and check out our Code of Conduct.
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Some of your past answers have not been well-received, and you're in danger of being blocked from answering.
Please pay close attention to the following guidance:
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3061827%2fcompute-int-0-pi-frac-cosnx-cosx-cosnt-cost-cosx-cost%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
2
Hello, welcome to MSE. This isn't a bad post as a new contributor, but you (and us) would benefit by adding some proofs for your 'testing' - show us your working for $n=1,2$. Good job on using MathJax!
– TheSimpliFire
Jan 4 at 16:56
@TheSimpliFire Added my working for n = 1, 2, thanks for the tip and the welcome., I need to calculate this integral to complete a problem from by problemset, which consists in calculating $int_0^{pi} dfrac{cos(nx) - cos(nt)}{cos(x) - cos(t)}dt$
– aleph0
Jan 4 at 17:13
Very nice solution (+1).
– TheSimpliFire
Jan 4 at 17:14
@aleph0 So you need to calculate$$intlimits_0^{pi}mathrm dt,frac {cos nx-cos nt}{cos x-cos t}$$?
– Frank W.
Jan 4 at 17:16
1
Are you sure that the result is correct?
– Zacky
Jan 4 at 17:41