# Math Help - Solving an Euler Equation (Calculus of variations)...

1. ## Solving an Euler Equation (Calculus of variations)...

I need to solve the Euler equation to make the following integral stationary:
$\int ( y'^2+\sqrt{y}) dx$
so I happily assume $F = y'^2+\sqrt{y}$ and find the partial derivative F wrt y' and F wrt y, put both into the Euler equation:
$\frac{d}{dx}(2y') -\frac{1}{2\sqrt{y}} = 0$

Route 1: simple rearrangement yield:
$y''- \frac{1}{4}y^{-\frac{1}{2}} = 0$

Route 2: intergate both side by x
$\int \sqrt{y} dy' = \frac{x}{4}$

Then I'm stuck in both.... there is probably a very simple way of getting the solution out, but my brain just can't turn the concer~ or I've already made a mistake early in my calculations?

Thanks for all the help!

2. ## Re: Solving an Euler Equation (Calculus of variations)...

Originally Posted by QDsolarX
I need to solve the Euler equation to make the following integral stationary:
$\int ( y'^2+\sqrt{y}) dx$
so I happily assume $F = y'^2+\sqrt{y}$ and find the partial derivative F wrt y' and F wrt y, put both into the Euler equation:
$\frac{d}{dx}(2y') -\frac{1}{2\sqrt{y}} = 0$

Route 1: simple rearrangement yield:
$y''- \frac{1}{4}y^{-\frac{1}{2}} = 0$

Route 2: intergate both side by x
$\int \sqrt{y} dy' = \frac{x}{4}$
This does not strike me as valid. You were good up to the point where you wrote down your second-order DE. I would go this route: whenever you have a DE of the form

$y''=f(y),$

you can multiply through by $y'$ and integrate at least once immediately. In your case, you get

$y''y'=\frac{1}{4}\,y^{-1/2}y'$

$\frac{(y')^{2}}{2}=\frac{\sqrt{y}}{2}+C_{1}.$

Now you have a first-order DE. Can you continue from here?

Then I'm stuck in both.... there is probably a very simple way of getting the solution out, but my brain just can't turn the concer~ or I've already made a mistake early in my calculations?

Thanks for all the help!