Graphing Infinite Series (Mathematica)

Eruditee
Messages
13
Reaction score
0

Homework Statement


I just have to graph this function to see where the "Gibbs phenomenon" occurs in its Fourier Series representation. I am pretty sure I integrated correctly.

Homework Equations


Fourier Series

The Attempt at a Solution


a_{0}=\frac{1}{\pi}\int_{-pi}^{pi}f(x)dx=\int_{-\pi}^{\pi}x^{2}dx=\frac{x^{3}}{3}|_{-\pi}^{\pi}=\frac{x^{3}}{3}|_{-\pi}^{-\frac{\pi}{2}}+\frac{x^{3}}{3}|_{-\frac{\pi}{2}}^{0}+\frac{x^{3}}{3}|_{0}^{\frac{\pi}{2}}+\frac{x^{3}}{3}|_{-\frac{\pi}{2}}^{\pi}==\frac{1}{\pi}(\frac{16\pi^{3}}{24})=\frac{2\pi^{2}}{3}<br />

With 2 successive integration by parts; I arrive at:
Fourier=\frac{\pi^{2}}{3}+4\overset{\infty}{\underset{1}{\sum}}\frac{cos(nx)}{n^{2}};cosine=even=\frac{\pi^{2}}{3}+4\overset{\infty}{\underset{1}{\sum}}\frac{(-1)^{n}}{n^{2}}

I do not know how to graph the first 5, 20, 400 etc terms. I have tried Mathematica, using
ListPlot[Table[{n, 4pi(-1)^n/(n^2)},{n,0,500}] but I get a blank graph. The problem is done; I just don't get how to graph this? I'd like to graph it from x= 0 to x= 4pi
 
Last edited by a moderator:
Physics news on Phys.org
You don't want to plot ##a_n##; you want to plot the approximations to f(x) using a finite number of terms to see Gibbs phenomenon.

Two reason you're getting a blank graph: First, you're dividing by zero. Second, Mathematica doesn't know what "pi" means. The case matters here. If you want the constant ##\pi##, you have to use Pi.
 
  • Like
Likes 1 person
vela said:
You don't want to plot ##a_n##; you want to plot the approximations to f(x) using a finite number of terms to see Gibbs phenomenon.

Two reason you're getting a blank graph: First, you're dividing by zero. Second, Mathematica doesn't know what "pi" means. The case matters here. If you want the constant ##\pi##, you have to use Pi.

Yes, I used Pi. I was just showing the actual forumula. Why would I not be able to plot 4cos(nx)/n^2 as (-1)/n^2, since I am asked to graph from 0 to 4 pi (cos 5pi = cos pi). It seems easier to keep it like the latter, in terms of n. I shound't start at zero, however; that's outside my series sum.

This is still blank"
ListPlot[Table[{n, x, (Pi)^2/(3)+ (4*cos (nx))/(n^2)}, {n, 1, 10}, {x, 0, 4 Pi}]]

Found this online; I need to default to youtube quicker,

I instead used the sum feature, like this down vote accepted


If you define

f[x_, t_, nm_] := 1/2 + Sum[(1/n/ Pi) ((-1)^n - 1) Sin[n x] Exp[-t n^2], {n, 1, nm}];

then

Plot[Table[f[x, t, 150], {t, {0, 0.01, 0.1, 0.5, 1, 10}}] // Release, {x, -Pi, Pi}]

with
f[x_, nm_] := ((Pi^2)/3) + Sum[(4 cos (nx)/(n^2)), {n, 1, nm}];
Plot[f[x_, nm_], {x, 0, 4 Pi}]
Manipulate[Plot[f[x_, nm_], {x, 0, 4 Pi}], {n, 1, 100}]Not working. Oh well. Next. It's not really a major part of the exercise as the calculation. This is harder than C IMO with all you have to know of it to make it servicable :-/. Thanx
 
Last edited by a moderator:
Does anyone else know what's going on? I have tried this several times, and am using a stackexchange dialog in which the following was adapted:

s[n_, x_] := ((Pi^2)/3) + Sum[((4 cos (kx))/(k^2)), {k, 1, n}]
partialsums = Table[s[n, x], {n, 1, 5}];
Plot[Evaluate[partialsums], {x, 0, 4 Pi}]

from
s[n_, x_] :=
8/4 + 3/(9 \[Pi]) Sum[(6 (-1)^k)/(k \[Pi]) Cos[(k \[Pi] x)/
2] + (16 (-1)^k + 13)/(\[Pi] k) Sin[(k \[Pi] x)/2], {k, 1, n}]

partialsums = Table[s[n, x], {n, 1, 5}];
Plot[Evaluate[partialsums], {x, -4, 4}]
http://mathematica.stackexchange.com/questions/2537/plotting-partial-sums-of-fourier-series

I still get a blank graph? Any help is appreciated as I do not have but a trial version of this software and like to finish this and my diff eq section.

Fixed **Spacing issue**
 
Last edited:
Your first line has a few errors. Use s[n_, x_] := ((Pi^2)/3) + Sum[((4 Cos[k x])/(k^2)), {k, 1, n}]
 
Whenever I cannot get a Plot the first thing I do is to look at the result of Table instead.

You already have the Table in partialsums. Look at that to see any clues for what went wrong.
With the previous hint about capitalization and with inserting a space or an * in the right place
I then get a plot that might be what you are looking for. I didn't even need the Evaluate.
(I think that using * instead of space probably saves time in the long run by avoiding errors)
 
Thank you both! I'm sorry I didn't make it more clear in the post: I had figured out it was a spacing issue. I haven't used that program in about 5 years, so I was a bit rusty.
 
Back
Top