1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    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!

Homework Help: Metric Spaces

  1. Dec 19, 2011 #1

    The sequential characterisation of continuity says that [itex]f[/itex] is continuous at [itex]x_0[/itex] if and only if for every sequence [itex](x_n)_{n\in\mathbb{N}}[/itex] in [itex]X[/itex], [itex]f(x_n)\to f(x_0)[/itex] as [itex]x_n \to x_0[/itex]. [itex]f[/itex] is continuous on [itex]X[/itex] if this is the case for all [itex]x_0 \in X[/itex].

    I think I've done all the parts of this question up to the last 2 parts.

    For part (b) is this right:

    Suppose [itex](x_n)_{n\in\mathbb{N}}[/itex] is a sequence in [itex]X[/itex] with [itex]x_n \to x\in X[/itex]. Then for all [itex]x\in X[/itex]: [tex]f(x_n) = (f_1(x_n) , f_2(x_n) , ... , f_N (x_n)) \to (f_1(x) , f_2(x) , ... , f_N (x) ) = f(x)[/tex] since all the [itex]f_i[/itex] are continuous.

    (This is also using a theorem which says that if [itex](x^{(n)})_{n\in\mathbb{N}}[/itex] is a sequence of vectors in [itex]\mathbb{R}^N[/itex] then [itex]x^{(n)} \to x\in\mathbb{R}^N[/itex] in the Euclidean metric [itex]\iff x_j^{(n)} \to x_j[/itex] for each [itex]1\leqslant j \leqslant N[/itex] in the standard metric on [itex]\mathbb{R}[/itex].)

    How would you show in the last 2 parts that [itex]F[/itex] and [itex]H[/itex] are continuous?
    Last edited: Dec 19, 2011
  2. jcsd
  3. Dec 20, 2011 #2
    EDIT: I see that [itex]F = \phi \circ f[/itex] and we've already shown that the composition of 2 continuous functions is continuous.

    What function can I compose [itex]F[/itex] with to turn the product in [itex]H[/itex] into a sum?
  4. Dec 21, 2011 #3
    Well, to turn a sum into a product we could use ##e^x## since ##e^{x+y} = e^x e^y ##. So what might you use to do the opposite?
  5. Dec 22, 2011 #4
    Last edited: Dec 22, 2011
  6. Dec 22, 2011 #5
    :smile: Right! And now this explains why the functions must be strictly positive for the last part.
  7. Dec 22, 2011 #6
    If [itex]h : (X,d_X) \to \mathbb{R}^N[/itex] is defined by [itex]h(x) = (h_1(x) , h_2(x) , ... , h_N(x) )[/itex] then [itex]h[/itex] is continuous by part (b).

    [itex]\displaystyle \log \circ H (x) = \log(H(x)) = \log \left( \prod_{j=1}^{N} h_j(x)^{a_j} \right)= \sum_{j=1}^N \log ( h_j(x)^{a_j} )= \sum_{j=1}^N a_j \log(h_j (x)) = \phi \circ (\log \circ h (x))[/itex]

    [itex]\phi , \log , h[/itex] are all continuous so their composition is continuous, so [itex]\log \circ H[/itex] is continuous.

    We know if [itex]f[/itex] and [itex]g[/itex] are 2 continuous functions then [itex]g \circ f[/itex] is continuous but [itex]g \circ f[/itex] continuous [itex]\not\Rightarrow f, g[/itex] are continuous so how do I frame the argument to show that H is continuous?
  8. Dec 22, 2011 #7


    User Avatar
    Science Advisor

    what continuous function can you compose with [itex]log \circ H[/itex] to recover H?
  9. Dec 22, 2011 #8
    [itex]\exp \circ ( \log \circ (H(x) ) = H(x) = \exp \circ (\phi \circ (\log \circ h(x)) )[/itex]

    [itex]\exp , \phi , \log , h[/itex] are all continuous so their compositions are all continuous, so H is continuous.
  10. Dec 22, 2011 #9


    User Avatar
    Science Advisor

    i'll buy that.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook