I am stuck on the exercise with integral equation

83 Views Asked by At

It's exercise 23.10 from Mathmatical Methods for Physics and Engineering (Riley K.F., Hobson M.P., Bence S.J. 3d Edition).

Show that the equation

$$f(x) = x^{-1/3} + \lambda\int\limits_{0}^{\infty}f(y)\exp(-xy)dy$$

has a solution of the form $Ax^{\alpha} + Bx^{\beta}$. Determine the values $\alpha$ and $\beta$, and show that those of $A$ and $B$ are $$\frac{1}{1-\lambda^2\Gamma\left(\frac{1}{3}\right)\Gamma\left(\frac{2}{3}\right)}\hspace{1cm}\text{and}\hspace{1cm}\frac{\lambda\Gamma\left(\frac{2}{3}\right)}{1-\lambda^2\Gamma\left(\frac{1}{3}\right)\Gamma\left(\frac{2}{3}\right)},$$ where $\Gamma(z)$ is the gamma fuction.

This is what I did: I wrote the Laplace transform of $f(x)$ as $\bar f(x)$ (I use the same letter for argument), so the original equation takes the form $$ f(x) = x^{-1/3} + \lambda \bar f(x).$$ Then I took the Laplace transform of that equation and got

$$ \bar f(x) = \Gamma\left(\frac{2}{3}\right)x^{-2/3} + \lambda \int\limits_{0}^{\infty} \frac{f(y)}{x + y}dy.$$ Then I substituted $\bar f(x)$ from the first equation and got $$f(x) = x^{-1/3} + \lambda\Gamma\left(\frac{2}{3}\right)x^{-2/3} + \lambda^2\int\limits_{0}^{\infty}\frac{f(y)}{x + y}dy.$$ And then I took the Laplace transform of this equation, so I got $$\bar f(x) = \Gamma\left(\frac{2}{3}\right)x^{-2/3} + \lambda\Gamma\left(\frac{2}{3}\right)\Gamma\left(\frac{1}{3}\right)x^{-1/3} + \lambda^2\int\limits_{0}^{\infty}f(y)\exp(xy)\Gamma(0,xy)dy.$$ Now I think that I did something wrong. But I can't understand what exactly that is.

1

There are 1 best solutions below

0
On BEST ANSWER

Since you have been given an ansatz, you can use it. Put $f(x) = Ax^\alpha + Bx^\beta$ back into the integral equation to get \begin{align*} Ax^\alpha + Bx^\beta = x^{-1/3} + \lambda\left(A\Gamma(\alpha+1)x^{-(\alpha+1)} + B\Gamma(\beta+1)x^{-(\beta+1)}\right) \end{align*}

Since you want this equation to hold for any $x>0$, you would want one of them, say $\alpha$, to be equal to $-1/3$ since we already have an $x^{-1/3}$ in the equation. $\beta$ would then have to be $-2/3$ because an $x^{-2/3}$ is introduced in the equation, and we don't get anything else. So now we can equate the coefficients of $x^{-1/3}$ and $x^{-2/3}$ on either side to get \begin{align*} A &= 1 + \lambda B\Gamma(1/3), \text{and}\\ B &= \lambda A\Gamma(2/3), \end{align*}

which gives \begin{align*} A &= \frac{1}{1-\lambda^2\Gamma(1/3)\Gamma(2/3)},\text{and}\\ B &= \frac{\lambda\Gamma(2/3)}{1-\lambda^2\Gamma(1/3)\Gamma(2/3)}. \end{align*}