1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Calculus of Variations: Functional is product of 2 integrals

  1. Oct 2, 2016 #1
    1. The problem statement, all variables and given/known data
    Minimize the functional: ∫01 dx y'2⋅ ∫01 dx(y(x)+1) with y(0)=0, y(1)=a

    2. Relevant equations
    (1) δI=∫ dx [∂f/∂y δy +∂f/∂y' δy']
    (2) δy'=d/dx(δy)
    (3) ∫ dx ∂f/∂y' δy' = δy ∂f/∂y' |01 - ∫ dx d/dx(∂f/∂y') δy
    where the first term goes to zero since there is no variation at the endpoints

    3. The attempt at a solution
    Using (1) to get
    δI= ∫dx 2y'δy' ∫dx(y+1) + ∫dx y'2 ∫dx δy
    Using (2) and (3) to get
    δI= ∫dx y'2 ∫dx δy - ∫dx d/dx(2y')δy ∫dx(y+1) = 0
    I then tried various ways of manipulating this equation to something that I could minimize trying to isolate δy but was unable to do so since it was always within an integral. I tried to simplify d/dx(2 y') to 2y'' to see if that helped but could not find any benefit to doing so. I also thought about using an integral by parts again but couldn't see where to do so. Using the values at the end points seems like something I should be doing but I don't know how I would. Any advice on how to proceed?
  2. jcsd
  3. Oct 3, 2016 #2


    User Avatar
    Staff Emeritus
    Science Advisor

    You're on the right track. You have:

    [itex]\int_0^1 dx (y')^2 \int_0^1 \delta y dx - \int_0^1 dx 2y'' \delta y\int_0^1 (y+1) dx = 0 [/itex]

    (Yes, it is better to write [itex]\frac{d}{dx} 2 y'[/itex] as [itex]2 y''[/itex].)

    At this point, let's just name two integrals:
    [itex]I_1 = \int_0^1 dx (y')^2[/itex]

    [itex]I_2 = \int_0^1 dx (y+1)[/itex]

    So we have:
    [itex]I_1 \int_0^1 \delta y dx - I_2 \int_0^1 dx 2y'' \delta y = 0[/itex]

    Now, the important thing to recognize is that [itex]I_1[/itex] and [itex]I_2[/itex] are not functions of [itex]x[/itex] (since [itex]x[/itex] has been integrated out). So you can treat them as if they were constants, and you can combine the two integrals:

    [itex]\int_0^1 \delta y (I_1 - I_2 2y'') dx = 0[/itex]
  4. Oct 3, 2016 #3
    Thank you! Yes, the integrals becoming constants after being evaluated was what I was failing to think about.
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted