I have the following functional

$ J(x)=\int\limits_{a}^{b}[{y(x)^2 + \dot{y}(x)^2 + 2 y(x) e^x}] dx $

and I want to find the extremum of it.

So I compute the Euler-Lagrange Equation and I get the result:

$ y^*(x) = c_1e^x+c_2e^{-x}-{e^x\over 4}+{xe^x\over 2} $

What can be done from here to find the $c_1$ and $c_2$ since I don't have any initial conditions ? Or at least the form that they should have ?



  1. Up to constant factors we may write OP's functional as $$ S[y]~:=~\int_a^b \!dx ~L , \qquad L ~:=~ \frac{1}{2}\dot{y}^2 + \frac{1}{2}( y+e^x)^2 ~\geq~0 .\tag{1}$$

  2. EL equation is $$ \ddot{y}-y~=~e^x.\tag{2}$$

  3. OP has already found the complete solution $$ y^{\ast}(x)~=~\left(\frac{x}{2}-\frac{3}{4}+c_+\right)e^x + c_-e^{-x},\tag{3}$$ so that $$\begin{align} \dot{y}^{\ast}(x)&~=~\left(\frac{x}{2}-\frac{1}{4}+c_+\right)e^x - c_-e^{-x}, \cr y^{\ast}(x)+e^x&~=~\left(\frac{x}{2}+\frac{1}{4}+c_+\right)e^x + c_-e^{-x}.\end{align}\tag{4}$$

  4. Therefore the Lagrangian becomes on-shell$^1$ equal to $$\begin{align}L^{\ast}(x)&~=~L^{\ast}_+(x)+L^{\ast}_-(x), \cr L^{\ast}_+(x)&~:=~ \left(\frac{x}{2}+c_+\right)^2e^{2x}~\geq~0, \cr L^{\ast}_-(x)&~:=~\left(\frac{1}{4}e^x+c_-e^{-x}\right)^2~\geq~0 ,\end{align}\tag{5}$$ while the functional on-shell is $$S[y^{\ast}]~=~S_+(c_+)+S_-(c_-), \qquad S_{\pm}(c_{\pm})~:=~\int_a^b \!dx ~L^{\ast}_{\pm}(x).\tag{6}$$

  5. Clearly it now just remains to minimize $S_{\pm}(c_{\pm})$ independently, which is a straightforward exercise.


$^1$ The word on-shell means that the EL eq. is satisfied.

  • $\begingroup$ "On shell" means that the argument of the Lagrangian solves the EL equations? $\endgroup$ – Giuseppe Negro Apr 10 at 11:20
  • $\begingroup$ $\uparrow$ Yes. $\endgroup$ – Qmechanic Apr 10 at 11:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.