Results 1 to 6 of 6

Math Help - Grade my continuity proof

  1. #1
    Member
    Joined
    Oct 2010
    Posts
    133

    Grade my continuity proof

    Prove that if f:R\rightarrow R is continuous and f(x+y)=f(x)+f(y) for all x,y\in R, then there is a contstant c\in R, such that f(x)=cx, for all x\in R.

    Here is my attempt:
    Proof:
    Let f:R\rightarrow R and f(x+y)=f(x)+f(y) for all x,y\in R, then there is a contstant c\in R, such that f(x)=cx, for all x\in R. Then, assume that f is NOT continuous. Then, suppose some point a is not a limit point for f. Then, since f(a)=ac, f is defined at a. Now,let a be a limit point for f. Then, 0<|x-a|<\delta \Longrightarrow |f(x)-f(a)|<\epsilon for some \epsilon>0.
    Then, 0<|x-a|<\delta \Longrightarrow |f(x)-f(a)|<\epsilon is the same as 0<|x-a|<\delta \Longrightarrow |cx-ca|<\epsilon.
    Now evaluate |x-a|<\delta. We have that |x-a|<|c||x-a|=|cx-ca|<\epsilon. Thus, \delta =\epsilon works. However, we have shown that if a is not a limit point at f then f is defined at a and that if a is a limit point of f then f(x)\longrightarrow f(a) as x\longrightarrow a. This is the definition of a continuous function so we have a contradiction. Thus, f is continuous.

    Any contructive critisism is welcome.
    Follow Math Help Forum on Facebook and Google+

  2. #2
    Member HappyJoe's Avatar
    Joined
    Sep 2010
    From
    Denmark
    Posts
    234
    Hello, zebra.

    Your proof does look somewhat confusing to me. I'll comment on parts of your proof.

    Quote Originally Posted by zebra2147 View Post
    Prove that if f:R\rightarrow R is continuous and f(x+y)=f(x)+f(y) for all x,y\in R, then there is a contstant c\in R, such that f(x)=cx, for all x\in R.

    Here is my attempt:
    Proof:
    Let f:R\rightarrow R and f(x+y)=f(x)+f(y) for all x,y\in R, then there is a contstant c\in R, such that f(x)=cx, for all x\in R.
    What do you mean exactly by this? You seem to immediately conclude that there is a constant c, such that f(x)=cx for all x, but this is what the whole problem is about provng.

    Quote Originally Posted by zebra2147 View Post
    Then, assume that f is NOT continuous.
    Why do you assume this? If you are going for either a proof by contraposition or a proof by contradiction, this is not the way to go.

    Quote Originally Posted by zebra2147 View Post
    Then, suppose some point a is not a limit point for f. Then, since f(a)=ac, f is defined at a. Now,let a be a limit point for f.
    Why did you start out by assuming that a is not a limit point of f, I mean, what do you use it for? You seem to conclude that f is then defined at a, but f is defined on all of the reals. And how do you know that f(a) = ac? Again it looks like you're using what you're actually supposed to prove.

    I'm afraid that what you do is based on wrong assumptions, i.e. you assume what you want to end up proving.
    Follow Math Help Forum on Facebook and Google+

  3. #3
    Banned
    Joined
    Oct 2009
    Posts
    4,261
    Thanks
    2
    Quote Originally Posted by zebra2147 View Post
    Prove that if f:R\rightarrow R is continuous and f(x+y)=f(x)+f(y) for all x,y\in R, then there is a contstant c\in R, such that f(x)=cx, for all x\in R.

    Here is my attempt:
    Proof:
    Let f:R\rightarrow R and f(x+y)=f(x)+f(y) for all x,y\in R, then there is a contstant c\in R, such that f(x)=cx, for all x\in R.


    Uh?? Why do you write the conclusion you must prove??


    Then, assume that f is NOT continuous.

    Why would you do that? This is a given, so you cannot assume it isn't true! You could negate the conclusion and try to prove by contradiction,

    but this ain't it.



    Then, suppose some point a is not a limit point for f. Then, since f(a)=ac


    This is wrong: you can't construct ANY prove assuming that what you HAVE to prove is correct!!


    , f is defined at a. Now,let a be a limit point for f. Then, 0<|x-a|<\delta \Longrightarrow |f(x)-f(a)|<\epsilon for some \epsilon>0.
    Then, 0<|x-a|<\delta \Longrightarrow |f(x)-f(a)|<\epsilon is the same as 0<|x-a|<\delta \Longrightarrow |cx-ca|<\epsilon.
    Now evaluate |x-a|<\delta. We have that |x-a|<|c||x-a|=|cx-ca|<\epsilon. Thus, \delta =\epsilon works. However, we have shown that if a is not a limit point at f then f is defined at a and that if a is a limit point of f then f(x)\longrightarrow f(a) as x\longrightarrow a. This is the definition of a continuous function so we have a contradiction. Thus, f is continuous.


    What a mess... ....this is NOT what you were asked to prove!!

    You really must read much more carefully the problem and understand it, and then attempt to solve it.

    Tonio



    Any contructive critisism is welcome.
    .
    Follow Math Help Forum on Facebook and Google+

  4. #4
    MHF Contributor Also sprach Zarathustra's Avatar
    Joined
    Dec 2009
    From
    Russia
    Posts
    1,506
    Thanks
    1
    Hint for the OP:

    Show that there exist c in R so that: f(x)=cx for all x in N(naturals), generalize it for all x in Z and then for all x in Q, conclude that f(x)=cx for all x in R. (Dense order)
    Follow Math Help Forum on Facebook and Google+

  5. #5
    Member
    Joined
    Oct 2010
    Posts
    133
    Here is attempt #2:
    Let's momentarily assume that the conclusion is correct.
    Consider: f(1) = 1c = c=f(1)

    Now, by induction, we can try to show that f(xn)=f(x)*n for any positive integer n, and any real number,x. We showed the base case above. Now, the induction step...

    f(x * (n + 1)) = f(x * n) + f(x) = f(x) * n + f(x) = f(x) * (n + 1)
    This shows that if we have a positive integer n, then:

    f(n) = f(1 * n) = f(1) * n = cn

    Then, we show that it works for negatives...
    f(-x) + f(x) = f(-x + x) = f(0) = 0<br />
So we have that, f(-x) = -f(x)

    Now we try to show that f(0) works too...
    f(0) = f(0 + 0) = f(0) + f(0)<br />
. So we have that f(0)=0

    We have shown that that for integers in f:
    f(n) = n * f(1)

    Using our results above we can look at rationals...
    f(p) = f(p/q * q)<br />
= f(p/q) * q <br />
f(p) / q = f(p/q)<br />
f(p/q) = (p / q)c

    So we have shown that when x is rational, f(x)=cx

    Now, we have to show that this works for all real numbers...

    First, suppose we have any real number x. Then, there exists a sequence of rationals, lets call it q(n), converging to that x. Above we showed that
    f(q(n)) = c * q(n) \longrightarrow cx
    However, since f is continuous, this means that cx = f(x). Therefore, for any x:
    f(x) = cx

    I think this is a better proof??? Although I know I kinda assumed the conclusion at the beginning. Any more feedback would be appreciated.
    Follow Math Help Forum on Facebook and Google+

  6. #6
    Banned
    Joined
    Oct 2009
    Posts
    4,261
    Thanks
    2
    Quote Originally Posted by zebra2147 View Post
    Here is attempt #2:
    Let's momentarily assume that the conclusion is correct.
    Consider: f(1) = 1c = c=f(1)

    Now, by induction, we can try to show that f(xn)=f(x)*n for any positive integer n, and any real number,x. We showed the base case above. Now, the induction step...

    f(x * (n + 1)) = f(x * n) + f(x) = f(x) * n + f(x) = f(x) * (n + 1)
    This shows that if we have a positive integer n, then:

    f(n) = f(1 * n) = f(1) * n = cn

    Then, we show that it works for negatives...
    f(-x) + f(x) = f(-x + x) = f(0) = 0<br />
So we have that, f(-x) = -f(x)

    Now we try to show that f(0) works too...
    f(0) = f(0 + 0) = f(0) + f(0)<br />
. So we have that f(0)=0

    We have shown that that for integers in f:
    f(n) = n * f(1)

    Using our results above we can look at rationals...
    f(p) = f(p/q * q)<br />
= f(p/q) * q <br />
f(p) / q = f(p/q)<br />
f(p/q) = (p / q)c

    So we have shown that when x is rational, f(x)=cx

    Now, we have to show that this works for all real numbers...

    First, suppose we have any real number x. Then, there exists a sequence of rationals, lets call it q(n), converging to that x. Above we showed that
    f(q(n)) = c * q(n) \longrightarrow cx
    However, since f is continuous, this means that cx = f(x). Therefore, for any x:
    f(x) = cx

    I think this is a better proof??? Although I know I kinda assumed the conclusion at the beginning. Any more feedback would be appreciated.

    Now the above almost begins to make sense. To begin with, f(0)=f(0+0)=f(0)+f(0)\Longrightarrow f(0)=0.

    Now, if n\in\mathbb{N} , then f(n)=f(1+...+1)=nf(1) (this can be shown this by induction, as you did), so putting f(1):=c we get that \forall n\in\mathbb{N}\,,\,\,f(n)=cn.

    Now take it from here just polishing up what you already did above.

    Tonio
    Follow Math Help Forum on Facebook and Google+

Similar Math Help Forum Discussions

  1. Grade my derivative proof
    Posted in the Differential Geometry Forum
    Replies: 8
    Last Post: January 18th 2011, 07:45 PM
  2. Grade My Uniform Continuity Proof
    Posted in the Differential Geometry Forum
    Replies: 2
    Last Post: November 8th 2010, 07:13 AM
  3. Challenging grade 6 question or grade 9 ?
    Posted in the Algebra Forum
    Replies: 2
    Last Post: June 24th 2010, 10:04 AM
  4. Replies: 1
    Last Post: June 6th 2010, 09:51 PM
  5. Proof to Grade (subsets)
    Posted in the Discrete Math Forum
    Replies: 4
    Last Post: September 16th 2007, 05:58 PM

Search Tags


/mathhelpforum @mathhelpforum