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

Derivatives question

  1. Feb 12, 2007 #1
    Hello, I am an economics student but this question is purely mathematical.

    Imagine the function U(x,y)=x^0.5y^0.5 ( sqrt(x) times sqrt(y) ) in space. If we then imagine the same function equal to a constant we get the 2d functions corresponding to the intersections of horizontal planes with the original 2 variable function, right? (if anyone here is taking economics, these would be the indifference curves). So that for every constant we get a function that is convex to the origin.
    Now imagine the function V=U^4=(x^2)(y^2) in space. If we do the same as above for this one, then for every constant we have a function that is concave to the origin, correct?
    The problem arises now. My professor says that, the derivative of the one variable functions (the convex and concave ones, the indifference curves) is the ratio of the partial derivatives of the original 2 variable function, which is always the same if the second is a monotonous increasing transformation of the first. Evidently, for the first one: (0.5x^-0.5y^0.5)/(0.5x^0.5y^-0.5) = y/x and for the second one 2xy^2/2x^2y = y/x. The problem is, if the slopes of the functions are the same for every point, then how come one is convex and the other concave? Can this be true? By drawing the curves I can see that it can, but only if we draw the curves that way. Why is it always true? I need to understand what the ratio of the partial derivatives of the original function means geometrically and logically, what does it represent?

    Last edited: Feb 12, 2007
  2. jcsd
  3. Feb 12, 2007 #2


    User Avatar
    Science Advisor
    Homework Helper

    What do you mean by "concave to the origin" and "convex to the origin"? Note the curve defined by V(x,y) = C for some (positive) constant C is symmetrical in the x-z and y-z planes. The restriction of this curve to the x > 0, y > 0 quadrant of space is THE EXACT SAME THING as the curve defined by U(x,y) = C1/4. So they're basically the same types of curves, and so I would assume their either both "concave to (0,0)" or both "convex to (0,0)".
  4. Feb 12, 2007 #3
    What I mean is, take the first function:
    U(x,y) = (x^0.5)(y^0.5)
    For any constant, for example 10: 10=(x^0.5)(y^0.5) <=> y=100/x is convex to (0,0).
    However, 10=(x^2)(y^2) <=> y=sqrt(10-x^2) is concave.
    These two different functions can not, as far as I can see, have the same slope for all x. However, both have the same ratio of partial derivatives in x over in y (y/x), which is supposed to be the slope of the curves of the original function for some constant. So how can this be true?
  5. Feb 13, 2007 #4


    User Avatar
    Science Advisor
    Homework Helper

    This is wrong.

    [tex]10 = x^2y^2 \Leftrightarrow y = \pm \frac{\sqrt{10}}{x}[/tex]

    I think you confused x2y2 with x2+y2.
  6. Feb 13, 2007 #5
    Thanks... I was thinking about it again on my way to school today and realised that, don't know how I did that... always a good idea to restart from the beginning..:redface: :shy:
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook