Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #6145 > unrolled thread

Simple DSolve equation

Started byRui <rui.rojo@gmail.com>
First post2011-11-03 08:52 +0000
Last post2011-11-12 12:41 +0000
Articles 4 — 2 participants

Back to article view | Back to comp.soft-sys.math.mathematica


Contents

  Simple DSolve equation Rui <rui.rojo@gmail.com> - 2011-11-03 08:52 +0000
    Re: Simple DSolve equation "Dr. Wolfgang Hintze" <weh@snafu.de> - 2011-11-04 10:59 +0000
    Re: Simple DSolve equation Rui <rui.rojo@gmail.com> - 2011-11-06 10:58 +0000
      Re: Simple DSolve equation "Dr. Wolfgang Hintze" <weh@snafu.de> - 2011-11-12 12:41 +0000

#6145 — Simple DSolve equation

FromRui <rui.rojo@gmail.com>
Date2011-11-03 08:52 +0000
SubjectSimple DSolve equation
Message-ID<j8tkpb$f14$1@smc.vnet.net>
Why does something like this not give the correct answer with restrictions over k?
How would you go about getting the right general solutions in these kind of basic differential equations?

Thanks

DSolve[{y''[x] == k y[x], y[0] == 0, y[10] == 0}, y[x], x]
Out={{y[x] -> 0}}

[toc] | [next] | [standalone]


#6171

From"Dr. Wolfgang Hintze" <weh@snafu.de>
Date2011-11-04 10:59 +0000
Message-ID<j90ghv$se0$1@smc.vnet.net>
In reply to#6145
"Rui" <rui.rojo@gmail.com> schrieb im Newsbeitrag 
news:j8tkpb$f14$1@smc.vnet.net...
> Why does something like this not give the correct answer with 
> restrictions over k?
> How would you go about getting the right general solutions in these 
> kind of basic differential equations?
>
> Thanks
>
> DSolve[{y''[x] == k y[x], y[0] == 0, y[10] == 0}, y[x], x]
> Out={{y[x] -> 0}}
>

Interesting question. Because I always had in mind to put down the 
procedure I'll do it here though the elaborated version might be a 
little lengthy. I have deleted the Ins and Outs and the error messages 
for ease of reading.


We will find all solutions (the trivial and he non-trivial ones) in the 
following way.
First we DSolve the differential equation with just one condition (at 
x= 0, say), and then impose the second condition as a Solve requirement 
on the set of solutions thus obtained.
We shall see that Solve does all the fine "quantization" we expect, i.e 
it gives us the possible values of the parameter k.
For better transparency we shall first apply Reduce, and Solve 
afterwards.

The code is found below

Best regards,
Wolfgang

Begin of code

f[x_] = y[x] /. DSolve[{Derivative[2][y][x] == k*y[x], y[0] == 0}, 
y[x], x][[1]]

(-E^((-Sqrt[k])*x))*(-1 + E^(2*Sqrt[k]*x))*C[2]

First Reduce,

r = Reduce[Simplify[f[10]] == 0]

(E^Sqrt[k] == -1 - E^(2*Sqrt[k]) - E^(3*Sqrt[k]) - E^(4*Sqrt[k]) && 1 + 
E^(2*Sqrt[k]) + E^(3*Sqrt[k]) + E^(4*Sqrt[k]) != 0) ||
(E^Sqrt[k] == 1 + E^(2*Sqrt[k]) - E^(3*Sqrt[k]) + E^(4*Sqrt[k]) && 1 + 
E^(2*Sqrt[k]) - E^(3*Sqrt[k]) + E^(4*Sqrt[k]) != 0) ||
(E^(2*Sqrt[k]) == -1 && E^Sqrt[k] != 0) || (E^(2*Sqrt[k]) == 1 + 
E^(4*Sqrt[k]) - E^(6*Sqrt[k]) + E^(8*Sqrt[k]) && E^Sqrt[k] != 0) ||
(C[2] == 0 && E^Sqrt[k] != 0) || E^Sqrt[k] == -1 || E^Sqrt[k] == 1

Length[r]
7

now Solve, one by one

n = 1;
r[[n]]
Solve[r[[n]]]
fs[n] = FullSimplify[f[x] /. %, x \[Element] Reals]

E^Sqrt[k] == -1 - E^(2*Sqrt[k]) - E^(3*Sqrt[k]) - E^(4*Sqrt[k]) && 1 + 
E^(2*Sqrt[k]) + E^(3*Sqrt[k]) + E^(4*Sqrt[k]) != 0

{{k -> -((16*Pi^2)/25)}, {k -> -((4*Pi^2)/25)}, {k -> 
Log[-(-1)^(1/5)]^2}, {k -> Log[-(-1)^(3/5)]^2}}

{-2*I*C[2]*Sin[(4*Pi*x)/5], -2*I*C[2]*Sin[(2*Pi*x)/5], -2*I*C[2]*Sin[(4*Pi*x)/5], 
 -2*I*C[2]*Sin[(2*Pi*x)/5]}


n = 2;
r[[n]]
Solve[r[[n]]]
fs[n] = FullSimplify[f[x] /. %, x \[Element] Reals]

E^Sqrt[k] == 1 + E^(2*Sqrt[k]) - E^(3*Sqrt[k]) + E^(4*Sqrt[k]) && 1 + 
E^(2*Sqrt[k]) - E^(3*Sqrt[k]) + E^(4*Sqrt[k]) != 0

{{k -> -((9*Pi^2)/25)}, {k -> -(Pi^2/25)}, {k -> Log[-(-1)^(2/5)]^2}, 
{k -> Log[-(-1)^(4/5)]^2}}

{-2*I*C[2]*Sin[(3*Pi*x)/5], -2*I*C[2]*Sin[(Pi*x)/5], -2*I*C[2]*Sin[(3*Pi*x)/5], 
 -2*I*C[2]*Sin[(Pi*x)/5]}


n = 3;
r[[n]]
Solve[r[[n]]]
fs[n] = FullSimplify[f[x] /. %, x \[Element] Reals]

E^(2*Sqrt[k]) == -1 && E^Sqrt[k] != 0

{{k -> -(Pi^2/4)}}

{-2*I*C[2]*Sin[(Pi*x)/2]}


n = 4;
r[[n]]
Solve[r[[n]]]
fs[n] = FullSimplify[f[x] /. %, x \[Element] Reals]

E^(2*Sqrt[k]) == 1 + E^(4*Sqrt[k]) - E^(6*Sqrt[k]) + E^(8*Sqrt[k]) && 
E^Sqrt[k] != 0

{{k -> -((81*Pi^2)/100)}, {k -> -((49*Pi^2)/100)}, 
{k -> -((9*Pi^2)/100)}, {k -> -(Pi^2/100)}, {k -> Log[-(-1)^(1/10)]^2},
{k -> Log[-(-1)^(3/10)]^2}, {k -> Log[-(-1)^(7/10)]^2}, {k -> 
Log[-(-1)^(9/10)]^2}}

{-2*I*C[2]*Sin[(9*Pi*x)/10], -2*I*C[2]*Sin[(7*Pi*x)/10], -2*I*C[2]*Sin[(3*Pi*x)/10], 
 -2*I*C[2]*Sin[(Pi*x)/10], -2*I*C[2]*Sin[(9*Pi*x)/10], -2*I*C[2]*Sin[(7*Pi*x)/10], 
 -2*I*C[2]*Sin[(3*Pi*x)/10], -2*I*C[2]*Sin[(Pi*x)/10]}


n = 5;
r[[n]]
Solve[r[[n]]]
fs[n] = FullSimplify[f[x] /. %, x \[Element] Reals]

C[2] == 0 && E^Sqrt[k] != 0

{0}


n = 6;
r[[n]]
Solve[r[[n]]]
fs[n] = FullSimplify[f[x] /. %, x \[Element] Reals]

E^Sqrt[k] == -1

{{k -> -Pi^2}}

{-2*I*C[2]*Sin[Pi*x]}


n = 7;
r[[n]]
Solve[r[[n]]]
fs[n] = FullSimplify[f[x] /. %, x \[Element] Reals]

E^Sqrt[k] == 1

{{k -> 0}}

{0}

Finally, we put everything together (and letting C[2] -> 1/(-2*I) which 
gives the list a nicer appeal)

Union[Flatten[Table[fs[i], {i, 1, 7}]]] /. C[2] -> 1/(-2*I)

{0, Sin[(Pi*x)/10], Sin[(Pi*x)/5], Sin[(3*Pi*x)/10], Sin[(2*Pi*x)/5], 
Sin[(Pi*x)/2], Sin[(3*Pi*x)/5], Sin[(7*Pi*x)/10], Sin[(4*Pi*x)/5], 
Sin[(9*Pi*x)/10], Sin[Pi*x]}

End of code

[toc] | [prev] | [next] | [standalone]


#6223

FromRui <rui.rojo@gmail.com>
Date2011-11-06 10:58 +0000
Message-ID<j95p9b$m6s$1@smc.vnet.net>
In reply to#6145
Wolfgang, but I think C[1] Sin[p/10 Pi x] are all solutions for p integer, right?

[toc] | [prev] | [next] | [standalone]


#6408

From"Dr. Wolfgang Hintze" <weh@snafu.de>
Date2011-11-12 12:41 +0000
Message-ID<j9lphi$lhm$1@smc.vnet.net>
In reply to#6223
"Rui" <rui.rojo@gmail.com> schrieb im Newsbeitrag 
news:j95p9b$m6s$1@smc.vnet.net...
> Wolfgang, but I think C[1] Sin[p/10 Pi x] are all solutions for p 
> integer, right?
>
>
Rui,

yes you are right, and these are all given in the list of results in my 
reply.

The point was that Mathematica indeed produces the correct Eigenvalues,
at least if you follow the procedure I have described.

The latter is, by the way, very close to what one nomally
does when solving Eigenvalue problems from diffenrential equations.

Wolfgang 

[toc] | [prev] | [standalone]


Back to top | Article view | comp.soft-sys.math.mathematica


csiph-web