Find Functions: f o g = Iℝ (ℝ→ℝ)

  • Thread starter Thread starter ver_mathstats
  • Start date Start date
  • Tags Tags
    Functions
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
21 replies · 2K views
ver_mathstats
Messages
258
Reaction score
21

Homework Statement


Find the functions:
f: (0, ∞) → ℝ and g: ℝ → ( 0, ∞) such that f o g = I (I denotes identity function on ℝ).

Homework Equations

The Attempt at a Solution


I am having trouble working backwards. I know that (f o g)(x) is f(g(x)). I am unsure if this is correct but would f o g be ℝ →ℝ be correct? And then that is the identity function? Making the identity function x?. I am confused on how exactly to approach this problem however. Any help is appreciated. Thank you.
 
on Phys.org
fresh_42 said:
I am confused, too, as there is no single solution, so "Find the functions" without further requirements doesn't make sense.
Can you at least name a few examples for ##f(g(x))=x\,?##
My apologies it's a homework question and I meant to put "find functions" rather than "find the functions", a few examples of f(g(x)) = x would be f(x) = x+4 and g(x) = x-4, and another example would be f(x) = x2 and g(x) = √x.
 
ver_mathstats said:
My apologies it's a homework question and I meant to put "find functions" rather than "find the functions", a few examples of f(g(x)) = x would be f(x) = x+4 and g(x) = x-4,

Can you see what's wrong with this example?

Hint: check all the criteria in the question.

Sorry, I misread the question.
 
Last edited:
I think the general solution is to take f(x) a random 1-1 function, and to take as ##g=f^{-1}## the inverse of f which is well defined since f is 1-1. So it will hold $$f(g(x))=f(f^{-1}(x))=x$$.
 
Delta2 said:
I think the general solution is to take f(x) a random 1-1 function, and to take as ##g=f^{-1}## the inverse of f which is well defined since f is 1-1. So it will hold $$f(g(x))=f(f^{-1}(x))=x$$.

That might not quite be correct either! See the example above.
 
Ok I think I see the problem, it has to do with the domain/image of the functions.

How about the functions ##g(x)=x^{2n}## ##f(x)={x}^{\frac{1}{2n}}## I think those two qualify all the criteria correct?
 
Delta2 said:
Ok I think I see the problem, it has to do with the domain/image of the functions.

How about the functions ##g(x)=x^2## ##f(x)=\sqrt{x}## I think those two qualify all the criteria correct?

Yes, of course.

The problem as stated is quite awkward, because you have to ensure that both functions are always positive. I'm not really sure what the point of the question is, unless it is to give a few examples.
 
Actually I am wrong cause in my example it is ##f(g(x))=|x|##... would work only if we consider ##g:R^{+}\to R^{+}## but the statement requires g to has domain R.

I think one point of the question is to understand that f,g must be some sort of inverse pair ##f,g=f^{-1}## but with the proper domains.

The only valid example I can think with well known functions is to take ##f(x)=\ln x##. You got in mind any other examples @PeroK?
 
Delta2 said:
Actually I am wrong cause in my example it is ##f(g(x))=|x|##... would work only if we consider ##g:R^{+}\to R^{+}## but the statement requires g to has domain R.

I think one point of the question is to understand that f,g must be some sort of inverse pair ##f,g=f^{-1}## but with the proper domains.

The only valid example I can think with well known functions is to take ##f(x)=\ln x##. You got in mind any other examples @PeroK?
##\ln x## isn't allowed either, as it's not always positive. Also, ##f## doesn't have to be 1-1, only ##g##. E.g. if ##g(x) = e^x##, then ##f(x) = \ln x## doesn't work, but you can still find a suitable ##f##.
 
fresh_42 said:
The easiest solution, i.e. which requires no thoughts about the domains, is probably a linear function. These are already ##\mathfrak{c}## many.
Wouldn't linear run into a problem with g>0?
How about just requiring g to be monotonic with an asymptotic value of zero in one direction and unbounded in the other?
 
  • Like
Likes   Reactions: Delta2
Well to be honest I am not sure . The statement of the problem shows as domain for g the whole ##\mathbb{R}##. I am not sure if that kind of statement implies that the domain can be a subset of ##\mathbb{R}##.
 
fresh_42 said:
Nothing what an absolute value can't fix.
It seems to fix one problem but creates another, then we will have ##f(g(x))=|x|## but the statement requires that ##f(g(x))=x## I made a similar mistake too in post #7,#9.
 
I obviously hadn't considered all aspects. Say ##M:=(0,\infty)##. Then we have ##f\, : \,M \longrightarrow \mathbb{R}## and ##g\, : \,\mathbb{R}\longrightarrow M## and ##f(g(x))=x## for all ##x\in \mathbb{R}##.

This means that ##g## is injective: ##g(a)=g(b) \Longrightarrow a=f(g(a))=f(g(b))=b##. So we need an embedding of the real numbers in ##M##. I think this has to be solved first.
 
  • Like
Likes   Reactions: Delta2