Tilted Forum Project Discussion Community  

Go Back   Tilted Forum Project Discussion Community > The Academy > Tilted Knowledge and How-To


 
 
LinkBack Thread Tools
Old 04-09-2005, 04:36 PM   #1 (permalink)
has a plan
 
Hain's Avatar
 
Location: middle of Whywouldanyonebethere
Rapidly finding terms of Taylor polynomial

w = INTEGRAL e<sup>sin(t)</sup> dt
w has no indefinite form. But a Taylor polynomial can be generated for
w(x) = INTEGRAL<sub>0</sub><sup>x</sup> e<sup>sin(t)</sup> dt
Where it is centered around zero. However, to do this rapidly there needs to be a quick way to determine the nth derivative of w(x).
I have noted that the nth derivative always has the term e<sup>sin(t)</sup>, so I have started with letting the nth derivative be equal to:
w<sup>(n)</sup>(x) = f<sub>n</sub>(x)*e<sup>sin(t)</sup>
Where:
f<sub>1</sub>(x) = 1 ;
f<sub>2</sub>(x) = cos(x) ;
f<sub>3</sub>(x) = cos<sup>2</sup>(x) - sin(x) ; so on and so forth if you wish to find each derivative.
If the derivative of w(x) continues then the pattern that f<sub>n</sub>(x) follows is:
f<sub>n</sub>(x) = f<sub>n-1</sub>(x)*cos(x) + <sup>d</sup>/<sub>dx</sub> f<sub>n-1</sub>(x)
Because the Taylor function we desire is centered at 0 let:
g(n) = f<sub>n</sub>(0)
The Taylor polynomial then becomes
SUM<sub>n=1</sub><sup>T</sup> g(n)*x<sup>n</sup>/n!
My question is: what is g(n) so that a single summation formula can be made to describe INTEGRAL<sub>0</sub><sup>x</sup> e<sup>sin(t)</sup> dt?
__________________

Last edited by Hain; 04-11-2005 at 07:55 AM.. Reason: Thank you CSFlim
Hain is offline  
Old 04-10-2005, 07:58 AM   #2 (permalink)
Sky Piercer
 
CSflim's Avatar
 
Location: Ireland
I'm don't know the answer to your question, but you can use the SUP and SUB html tags for superscript and subscript. i.e.
X[sup]y[/sup] gives: X<sup>y</sup>

when you replace the [] with <>
and

X[sub]n[/sub] gives: X<sub>n</sub>

you can even nest the tags if necessary:
X<sub>n</sub><sup>y<sub>i</sub><sup>2</sup></sup>
__________________
CSflim is offline  
Old 04-10-2005, 10:43 AM   #3 (permalink)
Junkie
 
Location: San Francisco
Sorry, the original version of my post didn't make any sense and I don't have time right now to make a new one that's helpful, but hopefully I'll figure something out later.
__________________
"Prohibition will work great injury to the cause of temperance. It is a species of intemperance within itself, for it goes beyond the bounds of reason in that it attempts to control a man's appetite by legislation, and makes a crime out of things that are not crimes. A Prohibition law strikes a blow at the very principles upon which our government was founded." --Abraham Lincoln

Last edited by n0nsensical; 04-10-2005 at 01:52 PM..
n0nsensical is offline  
Old 04-10-2005, 01:45 PM   #4 (permalink)
has a plan
 
Hain's Avatar
 
Location: middle of Whywouldanyonebethere
What I am trying to do is create a function that gives the numerical value of the nth derivative of e<sup>sin(t)</sup> at zero to create a single sum formula for the integral of that equation.

EDIT: Sorry, didn't see that. I really just woke up and am not banging on all cylinders as it were. Yes, it's 5 in the afternoon but I have been busy and have not had my morning dose of caffeine.
__________________

Last edited by Hain; 04-10-2005 at 01:55 PM..
Hain is offline  
Old 04-10-2005, 01:52 PM   #5 (permalink)
Junkie
 
Location: San Francisco
No, you misunderstand, I meant the original version of MY post didn't make sense and I edited it to say what it does now.
__________________
"Prohibition will work great injury to the cause of temperance. It is a species of intemperance within itself, for it goes beyond the bounds of reason in that it attempts to control a man's appetite by legislation, and makes a crime out of things that are not crimes. A Prohibition law strikes a blow at the very principles upon which our government was founded." --Abraham Lincoln
n0nsensical is offline  
Old 04-10-2005, 02:31 PM   #6 (permalink)
Addict
 
I don't have the answer, but to provide some data, here is the expansion of the derivatives of exp(sin(0)) for n=0,1,2,...,15:

1, 1, 1, 0, -3, -8, -3, 56, 217, 64, -2951, -12672, 5973, 309376, 1237173, -2917888.

The OEIS gives <a href="http://www.research.att.com/projects/OEIS?Anum=A002017">A002017</a>, which doesn't provide a general formula unfortunately.

If, in f<sub>n</sub>(t), we let s=sin(t) and c=cos(t), then the expansion looks like this:

f<sub>0</sub>(t) = 1s<sup>0</sup>c<sup>0</sup>.

f<sub>1</sub>(t) = 1s<sup>0</sup>c<sup>1</sup>.

f<sub>2</sub>(t) = -1s<sup>1</sup>c<sup>0</sup> + 1s<sup>0</sup>c<sup>2</sup>.

f<sub>3</sub>(t) = -1s<sup>0</sup>c<sup>1</sup> - 3s<sup>1</sup>c<sup>1</sup> + 1s<sup>0</sup>c<sup>3</sup>.

And so on. The coefficients here are trinomials. If I get any further I'll post here.

Last edited by phukraut; 04-10-2005 at 02:44 PM..
phukraut is offline  
Old 04-11-2005, 08:11 AM   #7 (permalink)
has a plan
 
Hain's Avatar
 
Location: middle of Whywouldanyonebethere
Phukraut, your system that you posted looks promising. But f<sub>1</sub>(x) = 1 by the way I am looking at w<sup>(n)</sup>(x).
w<sup>(n)</sup>(x) = <sup>d<sup>n-1</sup></sup>/<sub>dx<sup>n-1</sup></sub> e<sup>sin(x)</sup>
So e^sin(x) is th first deriviative.
__________________

Last edited by Hain; 04-11-2005 at 06:34 PM..
Hain is offline  
Old 04-11-2005, 08:49 AM   #8 (permalink)
Addict
 
Right. I'm not exactly sure why you've made the offset 1 instead of 0, but in the end the result is easily modified to account for this difference.
phukraut is offline  
Old 04-11-2005, 09:53 AM   #9 (permalink)
Addict
 
Looking at things a different way, we see a pattern. Let a<sub>ij</sub> be the coefficient in my above list for the term with sine to the power of i and cosine to the power of j---so for example, a<sub>32</sub>s<sup>3</sup>c<sup>2</sup>. Further, to avoid a later problem with notation, let n&gt;0. Now look at the values of f<sub>6</sub> and f<sub>7</sub> when we factor out cosine.

f<sub>6</sub>(t) = c<sup>0</sup>(a<sub>10</sub>s<sup>1</sup> + a<sub>20</sub>s<sup>2</sup> + a<sub>30</sub>s<sup>3</sup>) + c<sup>2</sup>(a<sub>02</sub>s<sup>0</sup> + a<sub>12</sub>s<sup>1</sup> + a<sub>22</sub>s<sup>2</sup>) + c<sup>4</sup>(a<sub>04</sub>s<sup>0</sup> + a<sub>14</sub>s<sup>1</sup>) + c<sup>6</sup>(a<sub>06</sub>s<sup>0</sup>).

f<sub>7</sub>(t) = c<sup>1</sup>(a<sub>01</sub>s<sup>0</sup> + a<sub>11</sub>s<sup>1</sup> + a<sub>21</sub>s<sup>2</sup> + a<sub>31</sub>s<sup>3</sup>) + c<sup>3</sup>(a<sub>03</sub>s<sup>0</sup> + a<sub>13</sub>s<sup>1</sup> + a<sub>23</sub>s<sup>2</sup>) + c<sup>5</sup>(a<sub>05</sub>s<sup>0</sup> + a<sub>15</sub>s<sup>1</sup>) + c<sup>7</sup>(a<sub>07</sub>s<sup>0</sup>).

I believe I have the pattern (if we let n&gt;0).

Let S<sub>m,p</sub> = a<sub>0p</sub>s<sup>0</sup> + a<sub>1p</sub>s<sup>1</sup> + ... + a<sub>mp</sub>s<sup>m</sup>, where m,p=0,1,2,.... Then we split n into even and odd terms and look at each pattern separately.

Consider k=1,2,3,..., and n=2k. Then we have



Then for odd n, (n=2k+1, where k=0,1,2,...), we see that



I hope I didn't make any mistakes there. Anyway, I'll try and combine the two later.

Last edited by phukraut; 04-11-2005 at 11:45 AM..
phukraut is offline  
Old 04-11-2005, 11:50 AM   #10 (permalink)
Addict
 
Taking t=0, we have



And for odd,



Now all that's left is to actually find the coefficients.

Last edited by phukraut; 04-11-2005 at 12:28 PM..
phukraut is offline  
Old 04-11-2005, 01:42 PM   #11 (permalink)
Addict
 
Since sin(0)=0, we can set a<sub>ij</sub>=0 for all i&gt;0, or simply disregard these coefficients. So here are some of the coefficients apparently solved as a<sub>0j</sub> (noting that binomial(i,j) is the binomial coefficient for integers i and j where, if i&lt;j, binomial(i,j)=0):

a<sub>0n</sub> = 1 for all n;
a<sub>0(n-2)</sub> = (-1)binomial(n,3) for all n;
a<sub>0(n-4)</sub> = (1/3)(5n - 22)binomial(n,5) for all n&gt;4.

Last edited by phukraut; 04-11-2005 at 03:19 PM.. Reason: Correction on the a<sub>ij</sub>=0. This is in fact not true. What I wanted to say was that we can disregard these coefficients since the term is identically zero.
phukraut is offline  
Old 04-11-2005, 02:38 PM   #12 (permalink)
has a plan
 
Hain's Avatar
 
Location: middle of Whywouldanyonebethere
I am blinking but it is starting to sink in. I have never dealt with this form of notation before. And the reason I am "offsetting" my derivatives is because the first derivative of w(x) = e^sin(x).
__________________
Hain is offline  
Old 04-11-2005, 03:10 PM   #13 (permalink)
Addict
 
This is just one approach. Another is to develop your differential above recursively. You may have more luck with that, since I'm currently stuch on the constants a<sub>0(n-6)</sub>, which begin (at n=7),

-1, -64, -820, -5440, -24970, -90112, -458172, -1036035, -1768195, -3935360, ....

Cheers.

Last edited by phukraut; 04-11-2005 at 03:20 PM..
phukraut is offline  
 

Tags
finding, polynomial, rapidly, taylor, terms


Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On



All times are GMT -8. The time now is 06:10 PM.

Tilted Forum Project

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2025, vBulletin Solutions, Inc.
Search Engine Optimization by vBSEO 3.6.0 PL2
© 2002-2012 Tilted Forum Project

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73