using the Fourier transform, solve











up vote
-1
down vote

favorite












using the Fourier transform, solve
$$dfrac{partial u}{partial t}=frac{partial^{2}u}{partial x^{2}},hspace{3mm}-infty<x<infty,hspace{2mm}t>0 \u(x,0)=left{begin{array}{lll}u_{0},&|x|<a\0,&|x|geq aend{array}right.hspace{4mm}|u(x,t)|<M$$










share|cite|improve this question






















  • Could you show some work? Where are you stuck?
    – Skip
    Aug 13 at 0:23










  • what happens is that I'm not sure what I did is okay, $$mathscr{F}left{frac{partial^{2}u}{partial x^{2}}right}=mathscr{F}left{frac{partial u}{partial t}right}$$ and I got that, $$mathscr{F}{u(x,0)}=int_{-infty}^{infty}f(x)e^{ialpha x}dx=u_{0}frac{e^{-ialpha}-e^{-ialpha}}{ialpha}$$ and I already got stuck
    – Santiago Seeker
    Aug 13 at 0:27

















up vote
-1
down vote

favorite












using the Fourier transform, solve
$$dfrac{partial u}{partial t}=frac{partial^{2}u}{partial x^{2}},hspace{3mm}-infty<x<infty,hspace{2mm}t>0 \u(x,0)=left{begin{array}{lll}u_{0},&|x|<a\0,&|x|geq aend{array}right.hspace{4mm}|u(x,t)|<M$$










share|cite|improve this question






















  • Could you show some work? Where are you stuck?
    – Skip
    Aug 13 at 0:23










  • what happens is that I'm not sure what I did is okay, $$mathscr{F}left{frac{partial^{2}u}{partial x^{2}}right}=mathscr{F}left{frac{partial u}{partial t}right}$$ and I got that, $$mathscr{F}{u(x,0)}=int_{-infty}^{infty}f(x)e^{ialpha x}dx=u_{0}frac{e^{-ialpha}-e^{-ialpha}}{ialpha}$$ and I already got stuck
    – Santiago Seeker
    Aug 13 at 0:27















up vote
-1
down vote

favorite









up vote
-1
down vote

favorite











using the Fourier transform, solve
$$dfrac{partial u}{partial t}=frac{partial^{2}u}{partial x^{2}},hspace{3mm}-infty<x<infty,hspace{2mm}t>0 \u(x,0)=left{begin{array}{lll}u_{0},&|x|<a\0,&|x|geq aend{array}right.hspace{4mm}|u(x,t)|<M$$










share|cite|improve this question













using the Fourier transform, solve
$$dfrac{partial u}{partial t}=frac{partial^{2}u}{partial x^{2}},hspace{3mm}-infty<x<infty,hspace{2mm}t>0 \u(x,0)=left{begin{array}{lll}u_{0},&|x|<a\0,&|x|geq aend{array}right.hspace{4mm}|u(x,t)|<M$$







real-analysis differential-equations fourier-transform






share|cite|improve this question













share|cite|improve this question











share|cite|improve this question




share|cite|improve this question










asked Aug 13 at 0:19









Santiago Seeker

678




678












  • Could you show some work? Where are you stuck?
    – Skip
    Aug 13 at 0:23










  • what happens is that I'm not sure what I did is okay, $$mathscr{F}left{frac{partial^{2}u}{partial x^{2}}right}=mathscr{F}left{frac{partial u}{partial t}right}$$ and I got that, $$mathscr{F}{u(x,0)}=int_{-infty}^{infty}f(x)e^{ialpha x}dx=u_{0}frac{e^{-ialpha}-e^{-ialpha}}{ialpha}$$ and I already got stuck
    – Santiago Seeker
    Aug 13 at 0:27




















  • Could you show some work? Where are you stuck?
    – Skip
    Aug 13 at 0:23










  • what happens is that I'm not sure what I did is okay, $$mathscr{F}left{frac{partial^{2}u}{partial x^{2}}right}=mathscr{F}left{frac{partial u}{partial t}right}$$ and I got that, $$mathscr{F}{u(x,0)}=int_{-infty}^{infty}f(x)e^{ialpha x}dx=u_{0}frac{e^{-ialpha}-e^{-ialpha}}{ialpha}$$ and I already got stuck
    – Santiago Seeker
    Aug 13 at 0:27


















Could you show some work? Where are you stuck?
– Skip
Aug 13 at 0:23




Could you show some work? Where are you stuck?
– Skip
Aug 13 at 0:23












what happens is that I'm not sure what I did is okay, $$mathscr{F}left{frac{partial^{2}u}{partial x^{2}}right}=mathscr{F}left{frac{partial u}{partial t}right}$$ and I got that, $$mathscr{F}{u(x,0)}=int_{-infty}^{infty}f(x)e^{ialpha x}dx=u_{0}frac{e^{-ialpha}-e^{-ialpha}}{ialpha}$$ and I already got stuck
– Santiago Seeker
Aug 13 at 0:27






what happens is that I'm not sure what I did is okay, $$mathscr{F}left{frac{partial^{2}u}{partial x^{2}}right}=mathscr{F}left{frac{partial u}{partial t}right}$$ and I got that, $$mathscr{F}{u(x,0)}=int_{-infty}^{infty}f(x)e^{ialpha x}dx=u_{0}frac{e^{-ialpha}-e^{-ialpha}}{ialpha}$$ and I already got stuck
– Santiago Seeker
Aug 13 at 0:27












1 Answer
1






active

oldest

votes

















up vote
1
down vote













I know this question is old now but perhaps somebody else may use it later too. I may rewrite the problem as
begin{cases}
displaystyle frac{partial u}{partial t} = frac{partial^2 u}{partial x^2} \
u(x,0) = u_0Big(h(x+a)-h(x-a)Big)
end{cases}

where $h(cdot)$ is the unit step and it is given that $|u|<M$ where $Minmathbb{R}$ is some bound on $u$.



Applying the transform as $mathcal{F}[u(x,t)]=U(lambda,t)$,



$$mathcal{F}big[u_t-u_{xx}big]=frac{dU}{dt}-(-ilambda)^2 U=0$$
and
$$mathcal{F}big[u(x,0)big] = u_0 int_{-infty}^infty big(h(x+a) - h(x-a)big)e^{ilambda x}dx = u_0 int_{-a}^a e^{ilambda x}dx.$$



Evaluating this transform directly as you did gives



$$
begin{align}
mathcal{F}big[u(x,0)big] &= u_0 frac{e^{ilambda a} - e^{-ilambda a}}{ilambda} \
&= u_0frac{coslambda a + isinlambda a - coslambda a + isinlambda a}{ilambda} \
&= frac{2u_0}{lambda}sin lambda a.
end{align}
$$



Alternatively, we can say that because the limits are symmetric and $sinlambda x$ is odd in $x$ and $cos lambda x$ is even in $x$,
$$mathcal{F}big[u(x,0)big] = 2u_0 int_0^a coslambda x dx = frac{2u_0}{lambda}sinlambda a.$$



So the problem becomes



begin{cases}
U'=-lambda^2 U \
displaystyle U(lambda,0) = frac{2u_0}{lambda}sinlambda a
end{cases}



Which has the solution
$$U(lambda, t) = frac{2u_0}{lambda}sinlambda a e^{-lambda^2 t}.$$



Now the solution in the time domain is simply the inverse transform of this expression.



$$u(x,t) = frac{1}{2pi} int_{-infty}^infty frac{2 u_0}{lambda}sinlambda a , e^{-lambda^2 t} e^{-i lambda x} dlambda$$



Multiplying by $a/a$ helps us see that this integral definitely exists.



$$u(x,t) = frac{u_0 a}{pi} int_{-infty}^infty frac{sin a lambda}{a lambda} e^{-lambda^2 t} e^{-i lambda x} dlambda$$



Under the same reasoning, $frac{sin a lambda}{a lambda} e^{-lambda^2 t}$ is even in $lambda$, so the $sin -lambda x$ term contributes nothing to the integral, and still $cos lambda x$ is even.



$$u(x,t) = frac{2u_0 a}{pi} int_0^infty frac{sin a lambda}{a lambda} cos x lambda , e^{-tlambda^2} , dlambda$$



Alternatively, the solution may be expressed as a convolution.



$$u(x,t)=2u_0a , frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda , * , frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} , dlambda $$



Evaluating the convolution involves evaluating each involved inverse Fourier transform independently.



To begin, let



$$
begin{align}
f(x) &= frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda \
&= frac{1}{pi}int_0^infty frac{sin alambda}{alambda}cos xlambda , dlambda \
&= frac{1}{2pi a}int_0^infty frac{sinlambda(a + x) + sinlambda(a - x)}{lambda} , dlambda \
&= frac{1}{2pi a}left((a + x)int_0^infty frac{sinlambda(a + x)}{lambda(a + x)} , dlambda + (a - x)int_0^infty frac{sinlambda(a - x)}{lambda(a - x)} , dlambdaright) \
&= frac{1}{2pi a}left((a + x)int_0^infty operatorname{sinc}lambda(a + x) , dlambda + (a - x)int_0^infty operatorname{sinc}lambda(a - x) , dlambdaright)
end{align}
$$



Knowing that $operatorname{sinc}y = operatorname{sinc}-y$ and



$$int_{0}^infty operatorname{sinc}by , dy = frac{pi}{2|b|},$$



we know the value of these integrals



$$
begin{align}
int_0^infty operatorname{sinc}lambda(a + x) , dlambda &=
begin{cases}
displaystyle -frac{pi}{2(a + x)}, &a + x < 0\
infty, &a + x = 0\
displaystyle frac{pi}{2(a + x)}, &a + x > 0\
end{cases} \
&= frac{pi}{2(a + x)}operatorname{sgn}(a + x) \
\
int_0^infty operatorname{sinc}lambda(a - x) , dlambda &= frac{pi}{2(a - x)}operatorname{sgn}(a - x)
end{align}
$$



So



$$f(x) = frac{operatorname{sgn}(a + x) + operatorname{sgn}(a - x)}{4a}.$$



Now let



$$g(x,t) = frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} dlambda = frac{1}{pi}int_0^infty e^{-t lambda^2} cos xlambda , dlambda.$$



Note that



$$
begin{align}
frac{partial g}{partial x}(x,t) &= -frac{1}{pi}int_0^infty lambda e^{-tlambda^2}sin xlambda , dlambda \
&= frac{1}{2pi t}int_0^infty sin xlambda , de^{-tlambda^2} \
&= frac{1}{2pi t}left(sin xlambda , e^{-tlambda^2}Bigvert_0^infty - int_0^infty xcos xlambda , e^{-tlambda^2} , dlambdaright) \
&= -frac{x}{2pi t}int_0^infty e^{-tlambda^2} cos xlambda , dlambda \
&= -frac{x}{2t}g(x,t).
end{align}
$$



Knowing that $e^{-t lambda^2} = e^{-t (-lambda)^2}$ and



$$int_{-infty}^infty e^{-by^2} , dy = sqrt{frac{pi}{b}},$$



we know $g(x,t)$ if we can solve



$$
begin{cases}
displaystyle frac{partial g}{partial x}(x,t) = -frac{x}{2t}g(x,t) \
displaystyle g(0,t) = frac{1}{2sqrt{pi t}}.
end{cases}
$$



The solution is readily obtainable as



$$g(x,t) = frac{e^{-frac{x^2}{4t}}}{2sqrt{pi t}}.$$



Then we have the solution



$$u(x,t) = 2 u_0 aint_{-infty}^infty f(s)g(x-s,t) , ds = frac{u_0}{4sqrt{pi t}}int_{-infty}^infty big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds.$$



If $|s| > a$, then the integrand is exactly $0$. Otherwise each $operatorname{sgn}$ function evaluates to $1$ except at $|s| = a$ where one is $1$ and the other is $0$.
$$
begin{align}
u(x,t) &= frac{u_0}{4sqrt{pi t}}int_{-a}^a big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds \
&= frac{u_0}{2sqrt{pi t}} int_{-a}^a e^{-frac{(x-s)^2}{4t}} , ds \
&= frac{u_0}{2}left(operatorname{erf}left(frac{a + x}{2sqrt{t}}right) + operatorname{erf}left(frac{a - x}{2sqrt{t}}right) right)
end{align}
$$



As $operatorname{erf}y$ is bounded, the solution $u(x,t)$ is bounded.



$hskip 1 in$ Heat Equation Solution






share|cite|improve this answer





















    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',
    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
    });


    }
    });














    draft saved

    draft discarded


















    StackExchange.ready(
    function () {
    StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2880854%2fusing-the-fourier-transform-solve%23new-answer', 'question_page');
    }
    );

    Post as a guest















    Required, but never shown

























    1 Answer
    1






    active

    oldest

    votes








    1 Answer
    1






    active

    oldest

    votes









    active

    oldest

    votes






    active

    oldest

    votes








    up vote
    1
    down vote













    I know this question is old now but perhaps somebody else may use it later too. I may rewrite the problem as
    begin{cases}
    displaystyle frac{partial u}{partial t} = frac{partial^2 u}{partial x^2} \
    u(x,0) = u_0Big(h(x+a)-h(x-a)Big)
    end{cases}

    where $h(cdot)$ is the unit step and it is given that $|u|<M$ where $Minmathbb{R}$ is some bound on $u$.



    Applying the transform as $mathcal{F}[u(x,t)]=U(lambda,t)$,



    $$mathcal{F}big[u_t-u_{xx}big]=frac{dU}{dt}-(-ilambda)^2 U=0$$
    and
    $$mathcal{F}big[u(x,0)big] = u_0 int_{-infty}^infty big(h(x+a) - h(x-a)big)e^{ilambda x}dx = u_0 int_{-a}^a e^{ilambda x}dx.$$



    Evaluating this transform directly as you did gives



    $$
    begin{align}
    mathcal{F}big[u(x,0)big] &= u_0 frac{e^{ilambda a} - e^{-ilambda a}}{ilambda} \
    &= u_0frac{coslambda a + isinlambda a - coslambda a + isinlambda a}{ilambda} \
    &= frac{2u_0}{lambda}sin lambda a.
    end{align}
    $$



    Alternatively, we can say that because the limits are symmetric and $sinlambda x$ is odd in $x$ and $cos lambda x$ is even in $x$,
    $$mathcal{F}big[u(x,0)big] = 2u_0 int_0^a coslambda x dx = frac{2u_0}{lambda}sinlambda a.$$



    So the problem becomes



    begin{cases}
    U'=-lambda^2 U \
    displaystyle U(lambda,0) = frac{2u_0}{lambda}sinlambda a
    end{cases}



    Which has the solution
    $$U(lambda, t) = frac{2u_0}{lambda}sinlambda a e^{-lambda^2 t}.$$



    Now the solution in the time domain is simply the inverse transform of this expression.



    $$u(x,t) = frac{1}{2pi} int_{-infty}^infty frac{2 u_0}{lambda}sinlambda a , e^{-lambda^2 t} e^{-i lambda x} dlambda$$



    Multiplying by $a/a$ helps us see that this integral definitely exists.



    $$u(x,t) = frac{u_0 a}{pi} int_{-infty}^infty frac{sin a lambda}{a lambda} e^{-lambda^2 t} e^{-i lambda x} dlambda$$



    Under the same reasoning, $frac{sin a lambda}{a lambda} e^{-lambda^2 t}$ is even in $lambda$, so the $sin -lambda x$ term contributes nothing to the integral, and still $cos lambda x$ is even.



    $$u(x,t) = frac{2u_0 a}{pi} int_0^infty frac{sin a lambda}{a lambda} cos x lambda , e^{-tlambda^2} , dlambda$$



    Alternatively, the solution may be expressed as a convolution.



    $$u(x,t)=2u_0a , frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda , * , frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} , dlambda $$



    Evaluating the convolution involves evaluating each involved inverse Fourier transform independently.



    To begin, let



    $$
    begin{align}
    f(x) &= frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda \
    &= frac{1}{pi}int_0^infty frac{sin alambda}{alambda}cos xlambda , dlambda \
    &= frac{1}{2pi a}int_0^infty frac{sinlambda(a + x) + sinlambda(a - x)}{lambda} , dlambda \
    &= frac{1}{2pi a}left((a + x)int_0^infty frac{sinlambda(a + x)}{lambda(a + x)} , dlambda + (a - x)int_0^infty frac{sinlambda(a - x)}{lambda(a - x)} , dlambdaright) \
    &= frac{1}{2pi a}left((a + x)int_0^infty operatorname{sinc}lambda(a + x) , dlambda + (a - x)int_0^infty operatorname{sinc}lambda(a - x) , dlambdaright)
    end{align}
    $$



    Knowing that $operatorname{sinc}y = operatorname{sinc}-y$ and



    $$int_{0}^infty operatorname{sinc}by , dy = frac{pi}{2|b|},$$



    we know the value of these integrals



    $$
    begin{align}
    int_0^infty operatorname{sinc}lambda(a + x) , dlambda &=
    begin{cases}
    displaystyle -frac{pi}{2(a + x)}, &a + x < 0\
    infty, &a + x = 0\
    displaystyle frac{pi}{2(a + x)}, &a + x > 0\
    end{cases} \
    &= frac{pi}{2(a + x)}operatorname{sgn}(a + x) \
    \
    int_0^infty operatorname{sinc}lambda(a - x) , dlambda &= frac{pi}{2(a - x)}operatorname{sgn}(a - x)
    end{align}
    $$



    So



    $$f(x) = frac{operatorname{sgn}(a + x) + operatorname{sgn}(a - x)}{4a}.$$



    Now let



    $$g(x,t) = frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} dlambda = frac{1}{pi}int_0^infty e^{-t lambda^2} cos xlambda , dlambda.$$



    Note that



    $$
    begin{align}
    frac{partial g}{partial x}(x,t) &= -frac{1}{pi}int_0^infty lambda e^{-tlambda^2}sin xlambda , dlambda \
    &= frac{1}{2pi t}int_0^infty sin xlambda , de^{-tlambda^2} \
    &= frac{1}{2pi t}left(sin xlambda , e^{-tlambda^2}Bigvert_0^infty - int_0^infty xcos xlambda , e^{-tlambda^2} , dlambdaright) \
    &= -frac{x}{2pi t}int_0^infty e^{-tlambda^2} cos xlambda , dlambda \
    &= -frac{x}{2t}g(x,t).
    end{align}
    $$



    Knowing that $e^{-t lambda^2} = e^{-t (-lambda)^2}$ and



    $$int_{-infty}^infty e^{-by^2} , dy = sqrt{frac{pi}{b}},$$



    we know $g(x,t)$ if we can solve



    $$
    begin{cases}
    displaystyle frac{partial g}{partial x}(x,t) = -frac{x}{2t}g(x,t) \
    displaystyle g(0,t) = frac{1}{2sqrt{pi t}}.
    end{cases}
    $$



    The solution is readily obtainable as



    $$g(x,t) = frac{e^{-frac{x^2}{4t}}}{2sqrt{pi t}}.$$



    Then we have the solution



    $$u(x,t) = 2 u_0 aint_{-infty}^infty f(s)g(x-s,t) , ds = frac{u_0}{4sqrt{pi t}}int_{-infty}^infty big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds.$$



    If $|s| > a$, then the integrand is exactly $0$. Otherwise each $operatorname{sgn}$ function evaluates to $1$ except at $|s| = a$ where one is $1$ and the other is $0$.
    $$
    begin{align}
    u(x,t) &= frac{u_0}{4sqrt{pi t}}int_{-a}^a big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds \
    &= frac{u_0}{2sqrt{pi t}} int_{-a}^a e^{-frac{(x-s)^2}{4t}} , ds \
    &= frac{u_0}{2}left(operatorname{erf}left(frac{a + x}{2sqrt{t}}right) + operatorname{erf}left(frac{a - x}{2sqrt{t}}right) right)
    end{align}
    $$



    As $operatorname{erf}y$ is bounded, the solution $u(x,t)$ is bounded.



    $hskip 1 in$ Heat Equation Solution






    share|cite|improve this answer

























      up vote
      1
      down vote













      I know this question is old now but perhaps somebody else may use it later too. I may rewrite the problem as
      begin{cases}
      displaystyle frac{partial u}{partial t} = frac{partial^2 u}{partial x^2} \
      u(x,0) = u_0Big(h(x+a)-h(x-a)Big)
      end{cases}

      where $h(cdot)$ is the unit step and it is given that $|u|<M$ where $Minmathbb{R}$ is some bound on $u$.



      Applying the transform as $mathcal{F}[u(x,t)]=U(lambda,t)$,



      $$mathcal{F}big[u_t-u_{xx}big]=frac{dU}{dt}-(-ilambda)^2 U=0$$
      and
      $$mathcal{F}big[u(x,0)big] = u_0 int_{-infty}^infty big(h(x+a) - h(x-a)big)e^{ilambda x}dx = u_0 int_{-a}^a e^{ilambda x}dx.$$



      Evaluating this transform directly as you did gives



      $$
      begin{align}
      mathcal{F}big[u(x,0)big] &= u_0 frac{e^{ilambda a} - e^{-ilambda a}}{ilambda} \
      &= u_0frac{coslambda a + isinlambda a - coslambda a + isinlambda a}{ilambda} \
      &= frac{2u_0}{lambda}sin lambda a.
      end{align}
      $$



      Alternatively, we can say that because the limits are symmetric and $sinlambda x$ is odd in $x$ and $cos lambda x$ is even in $x$,
      $$mathcal{F}big[u(x,0)big] = 2u_0 int_0^a coslambda x dx = frac{2u_0}{lambda}sinlambda a.$$



      So the problem becomes



      begin{cases}
      U'=-lambda^2 U \
      displaystyle U(lambda,0) = frac{2u_0}{lambda}sinlambda a
      end{cases}



      Which has the solution
      $$U(lambda, t) = frac{2u_0}{lambda}sinlambda a e^{-lambda^2 t}.$$



      Now the solution in the time domain is simply the inverse transform of this expression.



      $$u(x,t) = frac{1}{2pi} int_{-infty}^infty frac{2 u_0}{lambda}sinlambda a , e^{-lambda^2 t} e^{-i lambda x} dlambda$$



      Multiplying by $a/a$ helps us see that this integral definitely exists.



      $$u(x,t) = frac{u_0 a}{pi} int_{-infty}^infty frac{sin a lambda}{a lambda} e^{-lambda^2 t} e^{-i lambda x} dlambda$$



      Under the same reasoning, $frac{sin a lambda}{a lambda} e^{-lambda^2 t}$ is even in $lambda$, so the $sin -lambda x$ term contributes nothing to the integral, and still $cos lambda x$ is even.



      $$u(x,t) = frac{2u_0 a}{pi} int_0^infty frac{sin a lambda}{a lambda} cos x lambda , e^{-tlambda^2} , dlambda$$



      Alternatively, the solution may be expressed as a convolution.



      $$u(x,t)=2u_0a , frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda , * , frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} , dlambda $$



      Evaluating the convolution involves evaluating each involved inverse Fourier transform independently.



      To begin, let



      $$
      begin{align}
      f(x) &= frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda \
      &= frac{1}{pi}int_0^infty frac{sin alambda}{alambda}cos xlambda , dlambda \
      &= frac{1}{2pi a}int_0^infty frac{sinlambda(a + x) + sinlambda(a - x)}{lambda} , dlambda \
      &= frac{1}{2pi a}left((a + x)int_0^infty frac{sinlambda(a + x)}{lambda(a + x)} , dlambda + (a - x)int_0^infty frac{sinlambda(a - x)}{lambda(a - x)} , dlambdaright) \
      &= frac{1}{2pi a}left((a + x)int_0^infty operatorname{sinc}lambda(a + x) , dlambda + (a - x)int_0^infty operatorname{sinc}lambda(a - x) , dlambdaright)
      end{align}
      $$



      Knowing that $operatorname{sinc}y = operatorname{sinc}-y$ and



      $$int_{0}^infty operatorname{sinc}by , dy = frac{pi}{2|b|},$$



      we know the value of these integrals



      $$
      begin{align}
      int_0^infty operatorname{sinc}lambda(a + x) , dlambda &=
      begin{cases}
      displaystyle -frac{pi}{2(a + x)}, &a + x < 0\
      infty, &a + x = 0\
      displaystyle frac{pi}{2(a + x)}, &a + x > 0\
      end{cases} \
      &= frac{pi}{2(a + x)}operatorname{sgn}(a + x) \
      \
      int_0^infty operatorname{sinc}lambda(a - x) , dlambda &= frac{pi}{2(a - x)}operatorname{sgn}(a - x)
      end{align}
      $$



      So



      $$f(x) = frac{operatorname{sgn}(a + x) + operatorname{sgn}(a - x)}{4a}.$$



      Now let



      $$g(x,t) = frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} dlambda = frac{1}{pi}int_0^infty e^{-t lambda^2} cos xlambda , dlambda.$$



      Note that



      $$
      begin{align}
      frac{partial g}{partial x}(x,t) &= -frac{1}{pi}int_0^infty lambda e^{-tlambda^2}sin xlambda , dlambda \
      &= frac{1}{2pi t}int_0^infty sin xlambda , de^{-tlambda^2} \
      &= frac{1}{2pi t}left(sin xlambda , e^{-tlambda^2}Bigvert_0^infty - int_0^infty xcos xlambda , e^{-tlambda^2} , dlambdaright) \
      &= -frac{x}{2pi t}int_0^infty e^{-tlambda^2} cos xlambda , dlambda \
      &= -frac{x}{2t}g(x,t).
      end{align}
      $$



      Knowing that $e^{-t lambda^2} = e^{-t (-lambda)^2}$ and



      $$int_{-infty}^infty e^{-by^2} , dy = sqrt{frac{pi}{b}},$$



      we know $g(x,t)$ if we can solve



      $$
      begin{cases}
      displaystyle frac{partial g}{partial x}(x,t) = -frac{x}{2t}g(x,t) \
      displaystyle g(0,t) = frac{1}{2sqrt{pi t}}.
      end{cases}
      $$



      The solution is readily obtainable as



      $$g(x,t) = frac{e^{-frac{x^2}{4t}}}{2sqrt{pi t}}.$$



      Then we have the solution



      $$u(x,t) = 2 u_0 aint_{-infty}^infty f(s)g(x-s,t) , ds = frac{u_0}{4sqrt{pi t}}int_{-infty}^infty big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds.$$



      If $|s| > a$, then the integrand is exactly $0$. Otherwise each $operatorname{sgn}$ function evaluates to $1$ except at $|s| = a$ where one is $1$ and the other is $0$.
      $$
      begin{align}
      u(x,t) &= frac{u_0}{4sqrt{pi t}}int_{-a}^a big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds \
      &= frac{u_0}{2sqrt{pi t}} int_{-a}^a e^{-frac{(x-s)^2}{4t}} , ds \
      &= frac{u_0}{2}left(operatorname{erf}left(frac{a + x}{2sqrt{t}}right) + operatorname{erf}left(frac{a - x}{2sqrt{t}}right) right)
      end{align}
      $$



      As $operatorname{erf}y$ is bounded, the solution $u(x,t)$ is bounded.



      $hskip 1 in$ Heat Equation Solution






      share|cite|improve this answer























        up vote
        1
        down vote










        up vote
        1
        down vote









        I know this question is old now but perhaps somebody else may use it later too. I may rewrite the problem as
        begin{cases}
        displaystyle frac{partial u}{partial t} = frac{partial^2 u}{partial x^2} \
        u(x,0) = u_0Big(h(x+a)-h(x-a)Big)
        end{cases}

        where $h(cdot)$ is the unit step and it is given that $|u|<M$ where $Minmathbb{R}$ is some bound on $u$.



        Applying the transform as $mathcal{F}[u(x,t)]=U(lambda,t)$,



        $$mathcal{F}big[u_t-u_{xx}big]=frac{dU}{dt}-(-ilambda)^2 U=0$$
        and
        $$mathcal{F}big[u(x,0)big] = u_0 int_{-infty}^infty big(h(x+a) - h(x-a)big)e^{ilambda x}dx = u_0 int_{-a}^a e^{ilambda x}dx.$$



        Evaluating this transform directly as you did gives



        $$
        begin{align}
        mathcal{F}big[u(x,0)big] &= u_0 frac{e^{ilambda a} - e^{-ilambda a}}{ilambda} \
        &= u_0frac{coslambda a + isinlambda a - coslambda a + isinlambda a}{ilambda} \
        &= frac{2u_0}{lambda}sin lambda a.
        end{align}
        $$



        Alternatively, we can say that because the limits are symmetric and $sinlambda x$ is odd in $x$ and $cos lambda x$ is even in $x$,
        $$mathcal{F}big[u(x,0)big] = 2u_0 int_0^a coslambda x dx = frac{2u_0}{lambda}sinlambda a.$$



        So the problem becomes



        begin{cases}
        U'=-lambda^2 U \
        displaystyle U(lambda,0) = frac{2u_0}{lambda}sinlambda a
        end{cases}



        Which has the solution
        $$U(lambda, t) = frac{2u_0}{lambda}sinlambda a e^{-lambda^2 t}.$$



        Now the solution in the time domain is simply the inverse transform of this expression.



        $$u(x,t) = frac{1}{2pi} int_{-infty}^infty frac{2 u_0}{lambda}sinlambda a , e^{-lambda^2 t} e^{-i lambda x} dlambda$$



        Multiplying by $a/a$ helps us see that this integral definitely exists.



        $$u(x,t) = frac{u_0 a}{pi} int_{-infty}^infty frac{sin a lambda}{a lambda} e^{-lambda^2 t} e^{-i lambda x} dlambda$$



        Under the same reasoning, $frac{sin a lambda}{a lambda} e^{-lambda^2 t}$ is even in $lambda$, so the $sin -lambda x$ term contributes nothing to the integral, and still $cos lambda x$ is even.



        $$u(x,t) = frac{2u_0 a}{pi} int_0^infty frac{sin a lambda}{a lambda} cos x lambda , e^{-tlambda^2} , dlambda$$



        Alternatively, the solution may be expressed as a convolution.



        $$u(x,t)=2u_0a , frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda , * , frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} , dlambda $$



        Evaluating the convolution involves evaluating each involved inverse Fourier transform independently.



        To begin, let



        $$
        begin{align}
        f(x) &= frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda \
        &= frac{1}{pi}int_0^infty frac{sin alambda}{alambda}cos xlambda , dlambda \
        &= frac{1}{2pi a}int_0^infty frac{sinlambda(a + x) + sinlambda(a - x)}{lambda} , dlambda \
        &= frac{1}{2pi a}left((a + x)int_0^infty frac{sinlambda(a + x)}{lambda(a + x)} , dlambda + (a - x)int_0^infty frac{sinlambda(a - x)}{lambda(a - x)} , dlambdaright) \
        &= frac{1}{2pi a}left((a + x)int_0^infty operatorname{sinc}lambda(a + x) , dlambda + (a - x)int_0^infty operatorname{sinc}lambda(a - x) , dlambdaright)
        end{align}
        $$



        Knowing that $operatorname{sinc}y = operatorname{sinc}-y$ and



        $$int_{0}^infty operatorname{sinc}by , dy = frac{pi}{2|b|},$$



        we know the value of these integrals



        $$
        begin{align}
        int_0^infty operatorname{sinc}lambda(a + x) , dlambda &=
        begin{cases}
        displaystyle -frac{pi}{2(a + x)}, &a + x < 0\
        infty, &a + x = 0\
        displaystyle frac{pi}{2(a + x)}, &a + x > 0\
        end{cases} \
        &= frac{pi}{2(a + x)}operatorname{sgn}(a + x) \
        \
        int_0^infty operatorname{sinc}lambda(a - x) , dlambda &= frac{pi}{2(a - x)}operatorname{sgn}(a - x)
        end{align}
        $$



        So



        $$f(x) = frac{operatorname{sgn}(a + x) + operatorname{sgn}(a - x)}{4a}.$$



        Now let



        $$g(x,t) = frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} dlambda = frac{1}{pi}int_0^infty e^{-t lambda^2} cos xlambda , dlambda.$$



        Note that



        $$
        begin{align}
        frac{partial g}{partial x}(x,t) &= -frac{1}{pi}int_0^infty lambda e^{-tlambda^2}sin xlambda , dlambda \
        &= frac{1}{2pi t}int_0^infty sin xlambda , de^{-tlambda^2} \
        &= frac{1}{2pi t}left(sin xlambda , e^{-tlambda^2}Bigvert_0^infty - int_0^infty xcos xlambda , e^{-tlambda^2} , dlambdaright) \
        &= -frac{x}{2pi t}int_0^infty e^{-tlambda^2} cos xlambda , dlambda \
        &= -frac{x}{2t}g(x,t).
        end{align}
        $$



        Knowing that $e^{-t lambda^2} = e^{-t (-lambda)^2}$ and



        $$int_{-infty}^infty e^{-by^2} , dy = sqrt{frac{pi}{b}},$$



        we know $g(x,t)$ if we can solve



        $$
        begin{cases}
        displaystyle frac{partial g}{partial x}(x,t) = -frac{x}{2t}g(x,t) \
        displaystyle g(0,t) = frac{1}{2sqrt{pi t}}.
        end{cases}
        $$



        The solution is readily obtainable as



        $$g(x,t) = frac{e^{-frac{x^2}{4t}}}{2sqrt{pi t}}.$$



        Then we have the solution



        $$u(x,t) = 2 u_0 aint_{-infty}^infty f(s)g(x-s,t) , ds = frac{u_0}{4sqrt{pi t}}int_{-infty}^infty big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds.$$



        If $|s| > a$, then the integrand is exactly $0$. Otherwise each $operatorname{sgn}$ function evaluates to $1$ except at $|s| = a$ where one is $1$ and the other is $0$.
        $$
        begin{align}
        u(x,t) &= frac{u_0}{4sqrt{pi t}}int_{-a}^a big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds \
        &= frac{u_0}{2sqrt{pi t}} int_{-a}^a e^{-frac{(x-s)^2}{4t}} , ds \
        &= frac{u_0}{2}left(operatorname{erf}left(frac{a + x}{2sqrt{t}}right) + operatorname{erf}left(frac{a - x}{2sqrt{t}}right) right)
        end{align}
        $$



        As $operatorname{erf}y$ is bounded, the solution $u(x,t)$ is bounded.



        $hskip 1 in$ Heat Equation Solution






        share|cite|improve this answer












        I know this question is old now but perhaps somebody else may use it later too. I may rewrite the problem as
        begin{cases}
        displaystyle frac{partial u}{partial t} = frac{partial^2 u}{partial x^2} \
        u(x,0) = u_0Big(h(x+a)-h(x-a)Big)
        end{cases}

        where $h(cdot)$ is the unit step and it is given that $|u|<M$ where $Minmathbb{R}$ is some bound on $u$.



        Applying the transform as $mathcal{F}[u(x,t)]=U(lambda,t)$,



        $$mathcal{F}big[u_t-u_{xx}big]=frac{dU}{dt}-(-ilambda)^2 U=0$$
        and
        $$mathcal{F}big[u(x,0)big] = u_0 int_{-infty}^infty big(h(x+a) - h(x-a)big)e^{ilambda x}dx = u_0 int_{-a}^a e^{ilambda x}dx.$$



        Evaluating this transform directly as you did gives



        $$
        begin{align}
        mathcal{F}big[u(x,0)big] &= u_0 frac{e^{ilambda a} - e^{-ilambda a}}{ilambda} \
        &= u_0frac{coslambda a + isinlambda a - coslambda a + isinlambda a}{ilambda} \
        &= frac{2u_0}{lambda}sin lambda a.
        end{align}
        $$



        Alternatively, we can say that because the limits are symmetric and $sinlambda x$ is odd in $x$ and $cos lambda x$ is even in $x$,
        $$mathcal{F}big[u(x,0)big] = 2u_0 int_0^a coslambda x dx = frac{2u_0}{lambda}sinlambda a.$$



        So the problem becomes



        begin{cases}
        U'=-lambda^2 U \
        displaystyle U(lambda,0) = frac{2u_0}{lambda}sinlambda a
        end{cases}



        Which has the solution
        $$U(lambda, t) = frac{2u_0}{lambda}sinlambda a e^{-lambda^2 t}.$$



        Now the solution in the time domain is simply the inverse transform of this expression.



        $$u(x,t) = frac{1}{2pi} int_{-infty}^infty frac{2 u_0}{lambda}sinlambda a , e^{-lambda^2 t} e^{-i lambda x} dlambda$$



        Multiplying by $a/a$ helps us see that this integral definitely exists.



        $$u(x,t) = frac{u_0 a}{pi} int_{-infty}^infty frac{sin a lambda}{a lambda} e^{-lambda^2 t} e^{-i lambda x} dlambda$$



        Under the same reasoning, $frac{sin a lambda}{a lambda} e^{-lambda^2 t}$ is even in $lambda$, so the $sin -lambda x$ term contributes nothing to the integral, and still $cos lambda x$ is even.



        $$u(x,t) = frac{2u_0 a}{pi} int_0^infty frac{sin a lambda}{a lambda} cos x lambda , e^{-tlambda^2} , dlambda$$



        Alternatively, the solution may be expressed as a convolution.



        $$u(x,t)=2u_0a , frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda , * , frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} , dlambda $$



        Evaluating the convolution involves evaluating each involved inverse Fourier transform independently.



        To begin, let



        $$
        begin{align}
        f(x) &= frac{1}{2pi}int_{-infty}^inftyfrac{sin alambda}{alambda}e^{-i x lambda} , dlambda \
        &= frac{1}{pi}int_0^infty frac{sin alambda}{alambda}cos xlambda , dlambda \
        &= frac{1}{2pi a}int_0^infty frac{sinlambda(a + x) + sinlambda(a - x)}{lambda} , dlambda \
        &= frac{1}{2pi a}left((a + x)int_0^infty frac{sinlambda(a + x)}{lambda(a + x)} , dlambda + (a - x)int_0^infty frac{sinlambda(a - x)}{lambda(a - x)} , dlambdaright) \
        &= frac{1}{2pi a}left((a + x)int_0^infty operatorname{sinc}lambda(a + x) , dlambda + (a - x)int_0^infty operatorname{sinc}lambda(a - x) , dlambdaright)
        end{align}
        $$



        Knowing that $operatorname{sinc}y = operatorname{sinc}-y$ and



        $$int_{0}^infty operatorname{sinc}by , dy = frac{pi}{2|b|},$$



        we know the value of these integrals



        $$
        begin{align}
        int_0^infty operatorname{sinc}lambda(a + x) , dlambda &=
        begin{cases}
        displaystyle -frac{pi}{2(a + x)}, &a + x < 0\
        infty, &a + x = 0\
        displaystyle frac{pi}{2(a + x)}, &a + x > 0\
        end{cases} \
        &= frac{pi}{2(a + x)}operatorname{sgn}(a + x) \
        \
        int_0^infty operatorname{sinc}lambda(a - x) , dlambda &= frac{pi}{2(a - x)}operatorname{sgn}(a - x)
        end{align}
        $$



        So



        $$f(x) = frac{operatorname{sgn}(a + x) + operatorname{sgn}(a - x)}{4a}.$$



        Now let



        $$g(x,t) = frac{1}{2pi}int_{-infty}^infty e^{-t lambda^2} e^{-i x lambda} dlambda = frac{1}{pi}int_0^infty e^{-t lambda^2} cos xlambda , dlambda.$$



        Note that



        $$
        begin{align}
        frac{partial g}{partial x}(x,t) &= -frac{1}{pi}int_0^infty lambda e^{-tlambda^2}sin xlambda , dlambda \
        &= frac{1}{2pi t}int_0^infty sin xlambda , de^{-tlambda^2} \
        &= frac{1}{2pi t}left(sin xlambda , e^{-tlambda^2}Bigvert_0^infty - int_0^infty xcos xlambda , e^{-tlambda^2} , dlambdaright) \
        &= -frac{x}{2pi t}int_0^infty e^{-tlambda^2} cos xlambda , dlambda \
        &= -frac{x}{2t}g(x,t).
        end{align}
        $$



        Knowing that $e^{-t lambda^2} = e^{-t (-lambda)^2}$ and



        $$int_{-infty}^infty e^{-by^2} , dy = sqrt{frac{pi}{b}},$$



        we know $g(x,t)$ if we can solve



        $$
        begin{cases}
        displaystyle frac{partial g}{partial x}(x,t) = -frac{x}{2t}g(x,t) \
        displaystyle g(0,t) = frac{1}{2sqrt{pi t}}.
        end{cases}
        $$



        The solution is readily obtainable as



        $$g(x,t) = frac{e^{-frac{x^2}{4t}}}{2sqrt{pi t}}.$$



        Then we have the solution



        $$u(x,t) = 2 u_0 aint_{-infty}^infty f(s)g(x-s,t) , ds = frac{u_0}{4sqrt{pi t}}int_{-infty}^infty big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds.$$



        If $|s| > a$, then the integrand is exactly $0$. Otherwise each $operatorname{sgn}$ function evaluates to $1$ except at $|s| = a$ where one is $1$ and the other is $0$.
        $$
        begin{align}
        u(x,t) &= frac{u_0}{4sqrt{pi t}}int_{-a}^a big(operatorname{sgn}(a + s) + operatorname{sgn}(a - s)big)e^{-frac{(x-s)^2}{4t}} , ds \
        &= frac{u_0}{2sqrt{pi t}} int_{-a}^a e^{-frac{(x-s)^2}{4t}} , ds \
        &= frac{u_0}{2}left(operatorname{erf}left(frac{a + x}{2sqrt{t}}right) + operatorname{erf}left(frac{a - x}{2sqrt{t}}right) right)
        end{align}
        $$



        As $operatorname{erf}y$ is bounded, the solution $u(x,t)$ is bounded.



        $hskip 1 in$ Heat Equation Solution







        share|cite|improve this answer












        share|cite|improve this answer



        share|cite|improve this answer










        answered Nov 21 at 21:50









        Skip

        1,127214




        1,127214






























            draft saved

            draft discarded




















































            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.




            draft saved


            draft discarded














            StackExchange.ready(
            function () {
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2880854%2fusing-the-fourier-transform-solve%23new-answer', 'question_page');
            }
            );

            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







            Popular posts from this blog

            Le Mesnil-Réaume

            Ida-Boy-Ed-Garten

            web3.py web3.isConnected() returns false always