Copy wrong

- The other day I wanted to write f(x(1 + y)) on the board, but wrote something else instead.
- What did you write?
- f(x)(1 + f(y)).
- That is almost the same.
- Yes, I agree, and my students thought the slight change didn’t matter!
- Does it?
- What do you mean?
- Are there functions f for which the two give the same results for all x and y?
- I doubt it very much.

Problem source: Berkeley Math Circle Contest.

4 Responses to “Copy wrong”

  1. Sue VanHattum Says:

    The identity function: f(x)=x (which of course means that f(y)=y also). Both give x(1+y) as output. But that’s boring. I wonder if there’s anything more interesting.

  2. Richard Sabey Says:

    For convenience in notation, write a for f(1). Where I use n later, let it be an integer.

    Substituting x=1 and y=1 into equation 1:

    f(2) = f(1)(1 + f(1)) = a + a²

    Substituting x=1 and y=2 into equation 1:

    f(3) = f(1)(1 + f(2))
    = a(1 + a + a²)
    = a + a² + a³

    Substituting x=1 and y=3 into equation 1:

    f(4) = f(1)(1 + f(3))
    = a(1 + a + a² + a³)
    = a + a² + a³ + a^4 ___ [2]

    Substituting x=2 and y=1 into equation 1:

    f(4) = f(2)(1 + f(1))
    = (a + a²)(1 + a)
    = a + 2a² + a³ ___ [3]

    From equations 2 and 3,

    a + a² + a³ + a^4 = a + 2a² + a³
    a^4 = a²

    so a=0 or a=1 or a=-1.

    If f(1) = -1, then, substituting x=1 in equation 1,

    f(1+y) = -(1+f(y)) ___ [4]

    so

    f(2+y) = -(1+f(1+y)) = f(y)

    that is, f is periodic with period 2. In particular,

    f(2) = 0
    f(3) = -1 ___ [5]

    Substituting x=2, y=1/2 in equation 1,

    f(3) = 0

    contradicting equation 5. Thus f(1) cannot equal -1.

    If f(1) = 0, then, substituting x=1 in equation 1,

    f(1+y) = 0

    yielding f(x)=0 for all x.

    If f(1) = 1, then, substituting x=1 in equation 1,

    f(1+y) = 1 + f(y)

    determining f(x) for all x once f(x) has been defined on the interval (0, 1):

    f(x) = floor(x) + f(frac(x))

    (a result that can be proved by induction on floor(x)), where floor(x) is the greatest integer less than or equal to x, and frac(x) = x – floor(x).

    f(x) on the interval (0, 1) can’t be arbitrarily chosen, however. Substituting x=1/n, y=n-1 in equation 1:

    1 = f(1) = f(1/n) * n

    so f(1/n) = 1/n and f(x) = x for every x that is a multiple of 1/n. Thus f(x) = x for all rational x. If f must be continuous, f(x)=x for all x.

    To summarise, if f must be continuous, the only possibilities are f(x)=0 for all x and f(x)=x for all x.

  3. Richard Sabey Says:

    Where my equation 1 was Jan’s equation:

    f(x(1 + y)) = f(x)(1 + f(y)) ___ [1]

  4. Jan Nordgreen Says:

    The domain of f was not stated in the problem.

    Richard has shown that if the domain is R and is continuous f is rather boring.

    What if the domain is not the R? Can the domain have ten elements, one hundred, million and many?

Leave a Reply

How to use LaTeX in a comment.

You can add images to your comment by clicking here.