1. Not finding help here? Sign up for a free 30min 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!

A Proof that f:R^n->R is linear if

  1. Apr 21, 2009 #1
    1. The problem statement, all variables and given/known data

    Prove that if a map f: R^n -> R if f(0)=0 satisfies f differentiable, tf(x)=f(tx) for all x in R^n, then f is linear. (Prove additivity)



    2. Relevant equations



    3. The attempt at a solution

    So the first thing I tried was taking the derivative using the Matrix (or vector in this case)

    We have that f'(x+y) = f'(x)+f'(y) by the linearity of the derivative.

    Could one then take line integrals to get the desired equality (A path from 0 to x, a path from 0 to y), we should have, added together should enter a path from 0 to (x+y), so long as we are integrating from the same starting point (What's bothering me about this is it's not important f(0)=0, since we can just make the starting point of the path integral equal...)

    However we haven't defined the line integral in class, so regardless I'm thinking this is probably not what the instructor wants us to gain from the problem. Anyone have any insights about how I might approach this (Other Ideas I have had include, Showing that the partials are all constant (If they are all constant, than the original function which produced them must have contained only linear dependence on each of the n co-ordinates... that would imply the original vector field was linear together with f(0)=0 yes?)

    EDIT: So I explored the second Idea. If we differentiate (gradient) the second term we get that

    tf'(tx)=tf'(x)-> f'(tx)=f'(x), so we know that if homogeneity holds, f'(x) is constant, which shows that all of the partials are constant. If all the partials are constant, then in each co-ordinate, we have linearity (partial f by partial x_1 = c, so that f at most depended only on x_1 in its first power). f is however a scalar field though, so while I have convinced myself of the truth of this and drawn pictures, how might I rigorously arrive at the equality I need (That f(x_i+y_i)=f(x_i)+f(y_i) for all i, where x_i denotes the projection of x onto the vector x_i,

    EDIT: I think I got it.

    Define a function O(x,y)=f(x+y)-f(x)-f(y)

    Then O'(x,y)= f'(x+y)-f'(x)-f'(y), but since f'(x)=f'(y), f'(x+y)=f'(x)+f'(y) so that O'(x,y) = 0.

    Then, since f(0)=0, O(0,0)=0, so that O(x) must be zero everywhere.
     
    Last edited: Apr 21, 2009
  2. jcsd
  3. Apr 21, 2009 #2

    Mark44

    Staff: Mentor

    Part of your first sentence is missing. Prove that a map f:R^n --> R is what? From the work you show, I could probably figure out what's missing, but I would rather you tell me than put in that effort.
     
  4. Apr 21, 2009 #3
    Sorry, we need to prove it's linear. Which since we have homogeneity, amounts to showing f(x+y)=f(x)+f(y)
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook