Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Cauchy principal value

  1. Feb 15, 2014 #1
    The cauchy principal value formula is:


    But why ε→0⁺ in both terms? The correct wouldn't be ε→0⁻ in 1st term and ε→0⁺ in 2nd term? Like:

    [tex]\lim_{\varepsilon \to 0^-}\int_{a}^{c-\varepsilon}f(x)dx + \lim_{\varepsilon \to 0^+}\int_{c+\varepsilon}^{b}f(x)dx[/tex]

  2. jcsd
  3. Feb 15, 2014 #2


    User Avatar
    Homework Helper

    no if you do that c is in the interval
    we want to exclude c
  4. Feb 16, 2014 #3
    but if I define the superior limit in first integral like c+ε, so the expression below will be correct now?

    [tex]\lim_{\varepsilon \to 0^-}\int_{a}^{c+\varepsilon}f(x)dx + \lim_{\varepsilon \to 0^+}\int_{c+\varepsilon}^{b}f(x)dx[/tex]
  5. Feb 16, 2014 #4


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    There's no difference between adding a small -ve ε and subtracting a small +ve ε.

    ... although that definition is not equivalent as now you have two separate limits, so what you've defined is the improper integral as both limits must exist independently.
  6. Feb 17, 2014 #5


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    This is definitely wrong! The correct definition has been given in the posting by Jhenrique! The important point of the definition of the Cauchy PV is to leave out a tiny SYMMETRICAL "window" around the singularity and then make this window arbitrarily small.

    The difference can be demonstrated by a simple example. E.g., take the Cauchy principle value
    [tex]I=\text{PV} \int_{-1}^{1} \mathrm{d} x \frac{1}{x}.[/tex]
    Now the correct definition is
    [tex]I=\lim_{\epsilon \rightarrow 0^+} \left (\int_{-1}^{-\epsilon} \mathrm{d} x \frac{1}{x}+\int_{\epsilon}^1 \mathrm{d x} \frac{1}{x} \right ) = \ln \epsilon-\ln \epsilon=0.[/tex]
    If you try to take the limits of the two integrals separately, these limits do not even exist in this way!
  7. Feb 17, 2014 #6


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    With all due respect, I think you've got confused about who posted what. JHenrique posted an alternative defn of the CPV, which I pointed out was in fact the defn of an Improper Integral.

    You've really muddied the waters if you're saying JH is correct with his alternative definition.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook