Let F be the collection of all infinitely differentiable functions ψ defined on the region V = { ( x , y , z ) ∣ ∣ 1 ≤ x 2 + y 2 + z 2 ≤ 4 } such that ψ ( x , y , z ) = 3 whenever x 2 + y 2 + z 2 = 1 , and such that ψ ( x , y , z ) = 2 whenever x 2 + y 2 + z 2 = 4 . For any function ψ ∈ F , consider the volume integral I ( ψ ) = ∭ V { ( ∂ x ∂ ψ ) 2 + ( ∂ y ∂ ψ ) 2 + ( ∂ z ∂ ψ ) 2 } d x d y d z As ψ ranges over F , the smallest possible value of I ( ψ ) is equal to A π , where A is a positive integer. What is the value of A ?
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
Can we directly use divergence theorem. I got a wrong answer.
Log in to reply
I have used the divergence theorem to show that I is minimized at harmonic functions. I am not sure that a more direct use would be possible...
Ok I got it. I made a mistake.
This is a nice problem. Do you have some collection of problems which deal with maximal principles in differential equations and analytic functions. These are such beautiful topics. Some huge volumes of books are written on maximal principles in differential equations.
Log in to reply
An Introductiom to Lagrangian and Hamiltonian Mechanics is a pretty good book to learn calculus of variations and contains a proof that harmonic functions extremises the mean-square gradient average.
Log in to reply
Ummm...........The link is not working.........
Log in to reply
@Aaghaz Mahajan – Ah perhaps the free copy I linked to has been removed. You can still probably search up the book title.
This is nice book on partial differential equations by Jurgen jost. There are some books which talk only about maximal principles in differential equations.
By symmetry, I assume that ψ depends on r alone, so that the integral becomes ∫ V ( ψ ′ ( r ) ) 2 d V = 4 π ∫ 1 2 ( ψ ′ ( r ) ) 2 r 2 d r . Calculus of variations gives ψ ′ ′ ( r ) r + 2 ψ ′ ( r ) = 0 for the minimal integral so ψ ( r ) = 1 + r 2 for the given boundary conditions. The integral now comes out to be 8 π .
@Otto Bretscher, I am not science student. But i know calculus of variations and Euler-Lagrange equation. d ψ d F − d r d ( d ψ ′ d F ) = 0 is the Euler-Lagrange equation assumming that F ( r , ψ , ψ ′ ) . But your differential equation looks different It is ψ ′ ′ ( r ) r + 2 ψ ′ ( r ) = 0 . Would you explain how did you derive that?
Log in to reply
It's the Euler-Lagrange equation with F ( r , ψ , ψ ′ ) = ( ψ ′ ( r ) ) 2 r 2 .
A little knowledge of physics gives the solution in just one line. Think of ψ as the electrostatic potential. Then I ( ψ ) is simply the energy of the associated electrostatic field enclosed between the spherical shells at r = 1 and r = 2 . It is well-known that Coulomb's potential minimizes this energy , when the shells are held at fixed potentials. Hence, the optimal ψ in the spherical coordinates is of the form ψ ∗ ( r ) = r A + b . Supplying the boundary conditions ψ ∗ ( 1 ) = 3 , ψ ∗ ( 2 ) = 2 , the optimal ψ turns out to be ψ ∗ ( r ) = r 2 + 1 . Hence, I ( ψ ∗ ) = ∫ 1 2 ( − d r d ψ ∗ ( r ) ) 2 4 π r 2 d r = ∫ 1 2 ( r 2 2 ) 2 4 π r 2 d r = 8 π .
True, but physics is based on mathematical principles, and the fact that the Coulomb potential minimizes the energy of the associated electrostatic field between two equipotential shells follows from the fact that the Coulomb potential is a harmonic function (and my solution proves why this works).
Log in to reply
The solution that you provided is well-known and is available in many college-level texts. My solution shows that you can solve this problem even with high-school knowledge in Mathematics and Physics. It is about recognizing how to relate two apparently disparate things, which itself is a very useful exercise.
Problem Loading...
Note Loading...
Set Loading...
Suppose that ϕ , ψ are elements of F . Suppose moreover that ∇ 2 ϕ = 0 . By Green's Theorem 0 = ∮ ∂ V ( ϕ − ψ ) ∇ ϕ ⋅ d S = ∫ V ∇ ⋅ ( ( ϕ − ψ ) ∇ ϕ ) d τ = ∫ V ( ( ϕ − ψ ) ∇ 2 ϕ + ∇ ( ϕ − ψ ) ⋅ ∇ ϕ ) d τ = ∫ V ( ∣ ∣ ∇ ϕ ∣ ∣ 2 − ∇ ϕ ⋅ ∇ ψ ) d τ Here the surface integral ∫ ∂ V is the sum of the surface integral over the sphere centred at the origin with radius 2 (with outward-pointing normal) and the surface integral over the sphere centred at the origin with radius 1 (and inward pointing normal). This means that ∫ V ∇ ϕ ⋅ ∇ ψ = ∫ V ∣ ∣ ∇ ϕ ∣ ∣ 2 d τ and hence ∫ V ∣ ∣ ∇ ϕ − ∇ ψ ∣ ∣ 2 d τ = ∫ V ( ∣ ∣ ∇ ϕ ∣ ∣ 2 + ∣ ∣ ∇ ψ ∣ ∣ 2 − 2 ∇ ϕ ⋅ ∇ ψ ) d τ = ∫ V ∣ ∣ ∇ ψ ∣ ∣ 2 d τ − ∫ V ∣ ∣ ∇ ϕ ∣ ∣ 2 d τ so that I ( ψ ) = ∫ V ∣ ∣ ∇ ψ ∣ ∣ 2 d τ ≥ ∫ V ∣ ∣ ∇ ϕ ∣ ∣ 2 d τ = I ( ϕ ) Thus the smallest value of I ( ψ ) over all ψ ∈ F is obtained by choosing ψ to be harmonic. We can also use this argument to show that F contains at most one harmonic function.Thus we need to find the harmonic function ϕ for 1 ≤ r ≤ 2 such that ϕ = 3 when r = 1 and ϕ = 2 when r = 2 . In other words, we want ϕ = 1 + r 2 so that ∇ ϕ = − 2 r − 3 r , and hence I ( ϕ ) = ∫ V ∣ ∣ ∇ ϕ ∣ ∣ 2 d τ = 4 π ∫ 1 2 4 r − 4 r 2 d r = 8 π making the answer 8 .