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!

Family of Regular Curves

  1. Jan 16, 2014 #1
    1. The problem statement, all variables and given/known data
    Let [itex]\alpha : I =[a,b]→R^{2}[/itex] be a rgular curve parametrized by arc length [itex]s[/itex]. Let [itex]f:I→R[/itex] be a differentiable function with [itex]f(a)=f(b)=0[/itex]. For small values of [itex]\epsilon[/itex]

    [itex]\alpha_{\epsilon}:=\alpha (s) + \epsilon f(s) N(s)[/itex]

    defines a one parameter family of regular curves having endpoints [itex]\alpha(a)[/itex] and [itex]\alpha(b)[/itex]. The length of these curves defines a function

    [itex]L(\epsilon):=L_{a}^{b}(\alpha_{\epsilon})=\int_{a}^{b} \left|\frac{d}{ds}\alpha_{\epsilon}\right| ds[/itex]

    By differentiating under the integral sign, show that

    [itex]L'(0)=-\int_{a}^{b}k_{\alpha}f(s)ds[/itex]

    where [itex]k_{\alpha}[/itex] is the curvature of the function [itex]\alpha[/itex].

    3. The attempt at a solution

    My attempt at the solution is somewhat limited to purely differentiating the term inside the integrand. I merely want to check that I am approaching this the right way. What I have so far is the following. Since the term [itex]k_{\alpha}f(s)[/itex] is continuous we can differentiate under the integral sign. We have therefore,

    [itex]\frac{d}{ds}\alpha_{\epsilon}=\alpha'(s)+\epsilon [f'(s)N(s)+f(s)N'(s)][/itex]

    [itex]=\alpha'(s)+\epsilon [f'(s)N(s)+f(s)(-k_{\alpha}(s)T(s))][/itex]

    Where T(s) is the tangent to the curve [itex]\alpha[/itex]. Furthermore, this gives

    [itex]\alpha'(s)+\epsilon [f'(s)(\frac{1}{k_{\alpha}(s)}T'(s)+f(s)(-k_{\alpha}(s)\alpha'(s))][/itex]

    [itex]=\alpha'(s)+\epsilon [f'(s)(\frac{1}{k_{\alpha}(s)}\alpha''(s)+f(s)(-k_{\alpha}(s)\alpha'(s))][/itex]

    Now this is where it gets tricky. I want to show

    [itex]L'(0)=-\int_{a}^{b}k_{\alpha}f(s)ds[/itex]

    which implies that I look at the derivative of the function at [itex]\epsilon=0[/itex]. But how would one differentiate [itex]\left|\frac{d}{ds}\alpha_{\epsilon}\right|[/itex] with the "modulus" present?
    Either I am over complicating this or I have simply forgotten this sort of exercise from previous work. . . Suggestions?
     
  2. jcsd
  3. Jan 23, 2014 #2
    I realize what my problem was now. I took a wrong turn and now realize that I must next differentiate with respect to [itex]\epsilon[/itex] then take the limit as [itex]\epsilon[/itex] approaches zero. Furthermore I could get away with utilizing the definition of the inner product [itex]\left|\frac{d}{ds}\alpha_{\epsilon}(s)\right|=\sqrt{<\frac{d}{ds}\alpha_{\epsilon}(s),\frac{d}{ds}\alpha_{\epsilon}(s)>}[/itex].

    Following through with some differentiation on [itex]s[/itex] and then differentiation the inner product with respect to [itex]\epsilon[/itex]; things simplify quite a bit once the limit is taken. The result immediately pops out after some useful inner product applications. Now to write down a rigorous proof of this.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted



Similar Discussions: Family of Regular Curves
  1. Families of Curves (Replies: 0)

Loading...