[Date Index]
[Thread Index]
[Author Index]
Re: Boundary Value Problem
*To*: mathgroup at smc.vnet.net
*Subject*: [mg7059] Re: Boundary Value Problem
*From*: Troy.D.Goodson at jpl.nasa.gov
*Date*: Sat, 3 May 1997 22:04:51 -0400 (EDT)
*Organization*: Reference.Com Posting Service
*Sender*: owner-wri-mathgroup at wolfram.com
On 24 Apr 1997 05:04:15 -0400, "Rick A. Sprague" <sprague at egr.msu.edu> wrote:
> Hello,
>
> Would anybody know how to solve this problem? As a mathcad user for about
> 1 1/2 years, I have been trying to make the transition to MMA by doing
> every problem out of the mathcad manual to learn MMA equivalents. I have
> come to a problem, though, that seems to have stumped MMA. Can the
> following problem be solved with out writing a lengthy MMA program?
>
> y'''''[x]+y[x]==0
>
> Conditions
> y[0]==0
> y'[0]==7
>
> y[1]==1
> y'[1]==10
> y''[1]=5
This is a little late, and maybe someone else has solved it, but here is a solution from Matlab
The solution I get is
y(0)= 0
y'(0)= 7.0000
y''(0)= 12.2378
y'''(0)= 9.3279
y''''(0)= -39.5937
The trick is that this is a linear problem and be transformed to:
y1 = y
y2 = y'
y3 = y''
y4 = y'''
y5 = y''''
y5' = -y
so if x = Transpose[y1 y2 y3 y4 y5], then
dx/dt = Ax, where...
>> A=[0 1 0 0 0;
0 0 1 0 0;
0 0 0 1 0;
0 0 0 0 1;
-1 0 0 0 0]
>> A1=A(:,1:2)
A1 =
0 1
0 0
0 0
0 0
-1 0
>> B=exp(A);
% The solution to the system is the matrix exponential
% B = e^(A*t)
%
% The TPBVP can be solved if we break up B:
%
% B = [ B1 B2 ] (B1 is 5x2)
% [ . . .] (B2 is 3x3)
% [ . . .]
%
% we know y1(0), y2(0) but we need y3(0), y4(0), & y5(0)
% so...
% evaluate multiply the
% leftmost 5x2 portion of B with the 2x1 vector of known
% initial conditions
>> B1=B(:,1:2);
>> c=B1*[0; 7];
>> c1=c(1:3,:)
c1 =
19.0280
7.0000
7.0000
>> d1=[1; 10; 5];
>> e1=d1-c1
e1 =
-18.0280
3.0000
-2.0000
%
% now we can solve e1=B2*yo, where
% yo = Transpose[y3(0) y4(0) y5(0)]
%
>> B2=B(1:3,3:5)
B2 =
1.0000 1.0000 1.0000
2.7183 1.0000 1.0000
1.0000 2.7183 1.0000
>> y2=inv(B2)*e1;
>> y0=[0; 7; y2]
y0 =
0
7.0000
12.2378
9.3279
-39.5937
>> B*y0
%
% this shows that the solution matches the t=1
% boundary conditions
%
ans =
1.0000
10.0000
5.0000
-79.0611
-11.0280
Troy
http://www.csun.edu/~kg46825/TGoodson.html
--
Posted using Reference.COM http://www.reference.com
Browse, Search and Post Usenet and Mailing list Archive and Catalog.
InReference, Inc. accepts no responsibility for the content of this posting.
Prev by Date:
**Re: Is a solution possible to this exponential equation?**
Next by Date:
**Re: Animated Graphics**
Previous by thread:
**Re: Is a solution possible to this exponential equation?**
Next by thread:
**Re: Boundary Value Problem**
| |