Functional Equations - Andrei Jorza - MOP (Red)

Please bring this handout to lecture (no extra handouts available).
Please
look over the ideas presented in the beginning of each chapter, and if you
have time then please look at the problems.
Functional Equations
Lecturer - Andrei Jorza
Red Group
Functional equations are quite popular at the IMO and other international math olympiads
but there is no generalized method for solving them. However, there are quite a few ideas
that are very useful.
Ideas
1. Cauchy’s functional equation is the following: f : A −→ A so that f (x + y) =

f (x) + f (y) (this property is called additivity). For certain sets A and certain proper-
ties of f , the function turns out to be linear. And in any ring A, the linear function
satisfies Cauchy’s equation. Here are a few examples:
(a) If f : Q −→ Q then f = f (1)x.

(b) If f : R −→ R and f is increasing, continuous or positive on the set (0, ∞) then
f = f (1)x.
(c) If f (x2 ) = f 2 (x) then f = f (1)x because then f is positive on the positives.
(d) If f : R −→ R then f (q) = f (1)q for all rational numbers q.
2. When you get to a functional equation of type, e.g. f (x3 ) = f 3 (x) or f (x3 ) = x2 f (x)
a good trick is to replace x by x + y. If we have additivity then this is very useful.
Remember USAMO 2002 problem 4?
3. In functional equations try to get to relations of the type f ◦ f = x (called unipotent)

or f ◦ f = f (called idempotent).
4. In the case of idempotent functions we can see that for any x ∈ Imf we have that
f (x) = x. In this case the surjectivity of the function is important.
5. For unipotence a good trick is setting x to be f (x).
6. When you get to the relation f 2 (x) = x2 do not conclude that f = x or f = −x.
Rather, this implies that f (x) = x for x ∈ E and f (x) = −x for x ∈ R \ E. From here
on assume there are x, y so that f (x) = x, f (y) = −y.
7. In some problems, where the functional equation involves x and y if we set y = x then
we get an equation of type f (g(x)) = g(x). The problem will be solved if we can prove
that f has at most one fixed point.
1
Problems
Real Functions
1. Find all f : R −→ R so that for all real x, y we have f (xf (x) + f (y)) = f 2 (x) + y.
(BMO 2000)
2. Find all f : R −→ R so that for all real x, y we have f (x2 + f (y)) = f 2 (x) + y. (IMO
1992)
3. Find f : (0, ∞) −→ (0, ∞) so that for all positive real x, y we have f (xf (y)) = yf (x)
and when x tends to ∞ then f (x) tends to 0. (IMO 1983)
4. Find f : (−1, ∞) −→ (−1, ∞) so that for all real x and y in (−1, ∞) we have
f (x + f (y) + xf (y)) = y + f (x) + yf (x)
and f (x)/x is increasing on each of (−1, 0) and (0, ∞). (IMO 1994)
5. Find the function f : [0, ∞) −→ [0, ∞) so that f (2) = 0 and f (x) 6= 0 for x ∈ [0, 2).
Also for all nonnegative reals x, y we have f (xf (y))f (y) = f (x + y). (IMO 1986)
Integer Functions
Integer functional equations ususally have different methods of approach than the real
ones. We still have same ideas that we want to use, such as idempotence and unipotence.
However, sometimes we need to rely on the special properties of the integers (and rationals).
1. One very important property is the uniqueness of decomposition into prime factors.
This property also extends to rationals if we allow negative powers as well.
2. Another quite important method used is writing numbers in bases different from base
10. Most often this different base is base 2.
Problems
1. Find a bijection f : N −→ N so that for all natural m and n we have
f (3mn + m + n) = 4f (m)f (n) + f (m) + f (n)
(Shortlist 1996)
2. Prove that there exists a function f : Q∗+ −→ Q∗+ so that for all positive rationals x
and y we have
f (x)
f (xf (y)) =
y
(IMO 1993)
3. Prove that there are no functions f : N −→ N so that we have (f ◦ f )(x) = x + 1987

for all natural numbers x. (IMO 1987)
2
4. Let S be the set of non-negative integers. Find all functions f : S −→ S so that
f (m + f (n)) = f (f (m)) + f (n)
for all m, n ∈ S. (IMO 1996)
5. A function f : N∗ −→ N∗ satisfies the properties f (1) = 1, f (3) = 3 and
f (2n) = f (n)
f (4n + 1) = 2f (2n + 1) − f (n)
f (4n + 3) = 3f (2n + 1) − 2f (n)
Find the numbers of n ≤ 1988 so that f (n) = n. (IMO 1988)
3
Homework
1. Let f : R −→ R so that for any two real numbers x and y we have f (x+y) = f (x)+f (y)
and f (xy) = f (x)f (y). Find f .
Hint: Use the properties of the Cauchy functional equations.
2. Prove that there is no bijection from the positive integers to the nonnegative integers
so that for all positive integers m, n we have
f (mn) = f (m) + f (n) + 3f (m)f (n)
(BMO 1991)
Hint: Look at problem 1 from Integer Functions.
3. Let f : [0, 1) ∩ Q −→ [0, 1) ∩ Q so that if x < 1/2 then f (x) = f (2x)/4 and if x ≥ 1/2
then f (x) = (3 + f (2x − 1))/4.
Hint: Look at problem 5 from Integer Functions
4. Let a ∈ R. Find all f : R −→ R so that f (0) = 1/2 and for all real x, y we have
f (x + y) = f (x)f (a − y) + f (a − x)f (y). (BMO 1987)
Hint: Prove that the function is a constant one.
Extra Problems (Harder)
1. Find all f : R −→ R so that for all real x, y we have
f (x − f (y)) = f (f (y)) + xf (y) + f (x) − 1
(IMO 1999)
2. Find the smallest value of f (1998) when f : N −→ N and for all natural numbers m, n
we have
f (n2 f (m)) = mf 2 (n)
(Shortlist 1998)
3. Let f : N∗ −→ N∗ so that for all positive integers m, n we have
f (f (m) + f (n)) = m + n
Find all possible values of f (1988). (Shortlist 1988)

Functional Equations - Andrei Jorza - MOP (Red)

Uploaded by

Copyright:

Available Formats

Functional Equations - Andrei Jorza - MOP (Red)

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Functional Equations - Andrei Jorza - MOP (Red)

Uploaded by

Copyright:

Available Formats

Please bring this handout to lecture (no extra handouts available).

1. Cauchy’s functional equation is the following: f : A −→ A so that f (x + y) =

(a) If f : Q −→ Q then f = f (1)x.

3. In functional equations try to get to relations of the type f ◦ f = x (called unipotent)

5. For unipotence a good trick is setting x to be f (x).

f (x + f (y) + xf (y)) = y + f (x) + yf (x)

1. Find a bijection f : N −→ N so that for all natural m and n we have

f (3mn + m + n) = 4f (m)f (n) + f (m) + f (n)

3. Prove that there are no functions f : N −→ N so that we have (f ◦ f )(x) = x + 1987

f (m + f (n)) = f (f (m)) + f (n)

for all m, n ∈ S. (IMO 1996)

5. A function f : N∗ −→ N∗ satisfies the properties f (1) = 1, f (3) = 3 and

Find the numbers of n ≤ 1988 so that f (n) = n. (IMO 1988)

f (mn) = f (m) + f (n) + 3f (m)f (n)

Extra Problems (Harder)

1. Find all f : R −→ R so that for all real x, y we have

f (x − f (y)) = f (f (y)) + xf (y) + f (x) − 1

3. Let f : N∗ −→ N∗ so that for all positive integers m, n we have

Find all possible values of f (1988). (Shortlist 1988)

You might also like