1. PF Contest - Win "Conquering the Physics GRE" book! Click Here to Enter
    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!

Complex Analysis

  1. Mar 20, 2009 #1
    Let [itex]w_1,w_2 \in \mathbb{C}[/itex] and [itex]\gamma[/itex] be some smooth curve from [itex]w_1[/itex] to [itex]w_2[/itex].

    Find [itex]\int_{\gamma} e^{\sin{z}} \cos{z} dz[/itex]

    this is holomorphic on the entire copmlex plan so we can't use a residue theorem. furthermore, we can't assume [itex]\gamma[/itex] is a closed contour as we aren't told [itex]w_1=w_2[/itex] so it looks as if we're going to need to parameterise [itex]\gamma[/itex].

    but we don't know what [itex]\gamma[/itex] looks like. however we do know that any two point in the copmlex plane can be joined by a finite number of horizontal and vertical lines so if we use instead of [itex]\gamma[/itex] a contour [itex]\gamma_1 \cup \gamma_2[/itex]

    where [itex]\gamma_1[/itex] is horizontal and [itex]\gamma_2[/itex] is vertical. this is my thoughts so far but parameterising these was still going to be pretty difficult so i decided to check if im on the right lines or not. any advice?
  2. jcsd
  3. Mar 20, 2009 #2


    User Avatar
    Homework Helper
    Gold Member

    Re: Copmlex Analysis

    Hint: is this integral path independent? How can you tell?
  4. Mar 21, 2009 #3
    Re: Copmlex Analysis

    well the fundamental theorem of calculus gives that
    [itex]\gamma:[a,b] \rightarrow U[/itex] with [itex]\gamma(a)=\alpha,\gamma(b)=\beta[/itex]
    and [itex]f: U \rightarrow \mathbb{C}[/itex] holomorphic on the open set U then, if [itex]\exists F'=f[/itex],

    [itex]\int_{\gamma} f(z) dz = \int_{\gamma} F'(z) dz = \int_a^b F'(\gamma(t)) \gamma'(t) dt = \int_a^b (F \cdot \gamma(t))' dt = F \cdot \gamma(b) - F \cdot \gamma(a) = F(\beta)- F(\alpha)[/itex]

    i.e. if we can find such an F, the integral will depend only on the end point values and will therefore be path independent

    for our case [itex]F=e^{\sin{z}}[/itex] is such an F and so the answer is

    [itex]\int_{\gamma} e^{\sin{z}} \cos{z} dz = e^{\sin{w_2}} - e^{\sin{w_1}}=e^{\sin{w_2}-\sin{w_1}}[/itex]

    is that ok?

    is there an easier way to tell its path independent?

    also, can you recommend a good book that has plenty of worked exapmles on integrals like this (and stuff using Cauchy residue theorem etc)???
  5. Mar 21, 2009 #4
    Re: Copmlex Analysis

    also, how did you realise to argue about path independence?
  6. Mar 21, 2009 #5
    A holomorphic function defined on [tex]\mathbb{C}[/tex] always has an antiderivative (for example compute the Taylor series and integrate each term).

    As you pointed out, existence of an antiderivate implies independence of path for the integral.

    More generally, suppose [tex]f:D\to\mathbb{C}[/tex] is a holomorphic function and [tex]\alpha, \beta[/tex] are paths with the same endpoints. If [tex]\alpha[/tex] can be continuously deformed in D into [tex]\beta[/tex] while keeping the endpoints fixed, then the integral of f over the two paths is the same. This follows easily from Cauchy's integral theorem.
  7. Mar 21, 2009 #6


    User Avatar
    Homework Helper
    Gold Member

    Re: Copmlex Analysis

    Aside from your final answer being in grievous error....sure:wink:

    Like yyat said, whenever the integrand is holomorphic, the integral is path independent.

    The question wouldn't have made much sense if the integral depended on the path now would it?
  8. Mar 21, 2009 #7
    yeah. the final answer step was a BAD mistake.
    surely then we could answer by this means for any given function but here it's particularly useful because we were given no information about the path itself other than the endpoints
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Similar Threads - Complex Analysis Date
Prove that this function is holomorphic Yesterday at 11:45 PM
Cauchy Integration Formula Jan 13, 2018
Laurent series of z^2sin(1/(z-1)) Jan 11, 2018
Struggling with Cardano Jan 9, 2018