solve an Isoperimetric problem in calculus of variation using Maxima open source.

110 Views Asked by At

Kind help to solve the below problem via maxima program

Find the extremal of the functional $\int_0^1((y^{'})^{2} + x^2)dx$ under the constraint $\int_0^1 y^2 dx\,=\,2$ and having $y(0)\,=\,0$ and $y(1)\,=\,1.$ Using Maxima programming.

I tried Like the below code kind help not coming

kill(all)$
ratprint: false$
depends(f,[x,y])$
depends(g,[x,y])$
depends(y,x)$
declare(P,real)$
assume(P>0)$ (As positive or negative or zero error was giving)
z: diff(y,x);
f:z^2+x^2;
g:y^2;
h:f+P*g;
h1:diff(h,y);
h2:diff(h,z);
h3:h1-diff(h2,x);
h4:ode2(h3,y,x);
h5: bc2(h4,x=0,y=0,x=1,y=1);
hr:  trigrat(h5);
h6:integrate((rhs(hr))^2,x,0,1);
h7:solve(trigrat(h6)=2,P);
trigrat(rhs(h7[1]));

Need help

Answer expected from maxima

y= %k1*%e^(sqrt(l)*x)+%k2*%e^(-sqrt(l)*x)
1

There are 1 best solutions below

0
On

This problem can be handled easily without the help of a symbolic processor so calling

$$ f(x,y,y',\lambda) = (y')^2+x^2+\lambda y^2 $$

we have

$$ f_y-\left(f_{y'}\right)'=y''-\lambda y = 0 $$

and after solving the ODE we have

$$ y = c_1 e^{\sqrt \lambda x}+c_2 e^{-\sqrt\lambda x} $$

with the boundary and integral conditions we need to solve

$$ \begin{array}{rcl} c_1+c_2 & = & 0\\ c_1 e^{\sqrt \lambda}+c_2 e^{-\sqrt\lambda} & = & 1\\ \frac{e^{-\sqrt{\lambda }} \left(e^{\sqrt{\lambda }}-1\right) \left(c_1 e^{\sqrt{\lambda }}+c_2\right)}{\sqrt{\lambda }} & = & 2 \end{array} $$

thus obtaining the constants $c_1, c_2, \lambda$ Here making the substitution $c_2 =-c_1$ we have

$$ \begin{array}{rcl} c_1 & = & -c_2\\2 c_1 \cosh \left(\sqrt{\lambda }\right) & = & 1\\ \frac{2 c_1 \left(\cosh \left(\sqrt{\lambda }\right)-1\right)}{\sqrt{\lambda }} & = & 2 \end{array} $$

This last step should be concluded using an iterative process which is not explicit in the MAXIMA script.

I hope this helps.