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

Partial Derivatives of Power Functions

  1. Apr 17, 2012 #1
    For a function such as


    How would you find the partial derivative of w with respect to y or z? I've tried using basic logarithmic differentiation, but can't arrive at the correct answer. For reference, the correct answer is

  2. jcsd
  3. Apr 17, 2012 #2
    Do you know how to differentiate things like


    (with a constant) or


    to x?? This problem is exactly the same as the one you've given
  4. Apr 17, 2012 #3
    To differentiate y=ax I would use logarithmic differentiation.

    1) Take the ln of both sides: ln(y)=ln(ax)

    2) Using properties of logs, this equals: ln(y)=x*ln(a)

    3) Differentiating each side: y'/y=ln(a)

    4) Solving for y': y'=y*ln(a)

    5) Substituting for y: y'=ax*ln(a)

    The derivative of y=ex is simply y'=ex*x', where x'=1.

    Not sure how knowing these helps solve this problem though...here's my attempt:


    1) Taking ln of both sides: ln(w)=ln(5xy/z)

    2) Using properties of logs: ln(w)=(y/z)*ln(5x)

    However, ln(5xy/z) ≠ (y/z)*ln(5x)
  5. Apr 17, 2012 #4
    There are several ways you can get around this. One way is to take the logarithm of w/5 instead of w. Another is to use the fact that ln(ab) = ln(a) + ln(b). Yet a third way is to note that you already know how to differentiate h(y) = x^(y/z) and use the linearity of differentiation:in particular, the derivative of c*h(y) for any constant c is just c times the derivative of h(y).
  6. Apr 17, 2012 #5
    Thanks slider.

    I originally thought that ln(abx)=xln(ab), but apparently that's not true. Yet ln(ax)=xln(a). Why is this so?
  7. Apr 18, 2012 #6
    ln((ab)x)=xln(ab), but ln(a(bx))≠xln(ab). Does that make sense?
  8. Apr 18, 2012 #7
    Got it. Thanks!
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook