Scribd
Upload
2 views4 pages

Feqn Day3 July 19

Uploaded by

siddharthadas.bns
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
2 views4 pages

Feqn Day3 July 19

Uploaded by

siddharthadas.bns
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Functional Equations (Day 3)

Online Class 14 Sunday, July 19

Problem 1. Find all functions f : R → R such that f (x2 ) − f (y 2 ) = (x + y)(f (x) − f (y))
holds for all x, y ∈ R.
Solution: We observe that if f satisfies this functional equation, then so does f + c for any
constant c. Motivated from this, we can assume w.l.o.g. that f (0) = 0. To formalise it, let
us define g(x) = f (x) − f (0) and observe that g satisfies the same functional equation. But
for g, we have an additional information, namely g(0) = 0.
We have
g(x2 ) − g(y 2 ) = (x + y)(g(x) − g(y)) (1)

Putting y = 0 in equation (1) we get g(x2 ) = xg(x). We substitute this back into (1) to get

g(x) g(y) 
xg(x) − yg(y) = (x + y)(g(x) − g(y)) =⇒ = (for all x, y 6= 0
x y

The last equation tells us that g(x)/x is a constant function, i.e. g(x) = kx for all x 6= 0 for
some constant k (which is g(1)). Since g(0) = 0, so this also holds for x = 0.
Thus, g(x) = kx and hence f (x) = kx + c for some constants k and c. We observe that for
any value of k and c, this function indeed satisfies the given functional equation.

Problem 2. A function f : Z+ → Z+ is defined as follows:


f (1) = 1, f (3) = 3, f (2n) = f (n),
f (4n + 1) = 2f (2n + 1) − f (n),
f (4n + 3) = 3f (2n + 1) − 2f (n).
Find the number of positive integers n ≤ 2020 for which f (n) = n.
Solution: First we check small values of n and hence make the following table.

n f(n) n in binary f(n) in binary


1 1 1 1
3 3 11 11
5 5 101 101
7 7 111 111
9 9 1001 1001
11 13 1011 1101
13 11 1101 1011
15 15 1111 1111
19 25 10011 11001
.. .. .. ..
. . . .

1
So here is our guess: if n = a0 + a1 2+ · · · + ak 2k = (ak · · · a1 a0 )2 be the binary representation
of n, then f (n) = (a0 a1 · · · ak )2 = a0 2k + a1 2k−1 + · · · + ak−1 2 + ak .
To prove this, we shall induct on k (or equivalently, on the number of digits of n). When
k = 0 or 1, we have checked that our claim holds. Next, suppose the claim holds for every
k < m. We shall now prove our claim for k = m.
Case 1: a0 = 0 (i.e., all even numbers with m + 1 digits in binary)
Since n = (am am−1 · · · a1 0)2 = 2 × (am am−1 · · · a1 )2 , we can say that

f (n) = f (2 × (am am−1 · · · a1 )2 ) = f ((am am−1 · · · a1 )2 ) = (a1 a2 · · · am )2 = (0a1 a2 · · · am )2

Case 2: a0 = 1 (i.e., all odd numbers with m + 1 digits in binary)


Subcase 1: a1 = 0, a0 = 1
This subcase is about all n of the form 4m + 1.
Let n = (am am−1 · · · a2 01)2 = 4 × (am am−1 · · · a2 )2 + 1. Hence
f (n) = f (4 × (am am−1 · · · a2 )2 + 1)
= 2f (2 × (am am−1 · · · a2 )2 + 1) − f ((am am−1 · · · a2 )2 )
= 2f ((am am−1 · · · a2 1)2 ) − f ((am am−1 · · · a2 )2 )
= 2(1a2 · · · am−1 am )2 − (a2 · · · am−1 am )2
= 2 × 2m−1 + (a2 · · · am−1 am )2
= (10a2 · · · am−1 am )2 .
Subcase 2: a1 = 1, a0 = 1
This subcase is about all n of the form 4m + 3.
Let n = (am am−1 · · · a2 11)2 = 4 × (am am−1 · · · a2 )2 + 3. Hence
f (n) = f (4 × (am am−1 · · · a2 )2 + 3)
= 3f (2 × (am am−1 · · · a2 )2 + 1) − 2f ((am am−1 · · · a2 )2 )
= 3f ((am am−1 · · · a2 1)2 ) − 2f ((am am−1 · · · a2 )2 )
= 3(1a2 · · · am−1 am )2 − 2(a2 · · · am−1 am )2
= 3 × 2m−1 + (a2 · · · am−1 am )2
= (2 + 1) × 2m−1 + (a2 · · · am−1 am )2
= (11a2 · · · am−1 am )2 .
This completes our induction and hence induction completes the proof of our claim.
Next we have to find the number of n ≤ 2020 such that f (n) = n. Note that f (n) = n holds
iff the binary representation of n is a palindrome. For any n ≤ 2020 < 211 , the number of
digits in binary representation of n will be ≤ 11.
For each 1 ≤ k ≤ 11, let us find the number of k digit binary numbers that are palindromic.
If k = 2m then this count will be 2m−1 and if k = 2m + 1 then it will be 2m .
Summing up, we get 5m=1 2m−1 + 5m=0 2m . Note that this is the number of such numbers
P P

n that are less than 2048. So we have to subtract the count of those n that are > 2020.

2
Since 2020 = (11111100100)2 so there is only one n > 2020 which is palindromic, namely
(11111111111)2 = 2047. Final answer: 93.

Problem 3. Let f : N → N be a function such that f (n + 1) > f (f (n)) for every n ∈ N.


Prove that f (n) = n for all n ∈ N.
Solution: We shall make use of the Well-Ordering Principle (WOP), which states that any
non-empty subset of N must have a smallest element. Let us begin with S1 = {f (n) : n ≥ 1}.
By WOP, it has a smallest element, say f (a1 ). We shall show that a1 = 1. If a1 > 1 then
f (a1 − 1) is well-defined and f (f (a1 − 1)) < f (a1 ) which contradicts the minimality of f (a1 ).
Therefore, we conclude a1 = 1. Is it possible that f (1) = f (n) for some n > 1? If so, then
f (1) = f (n) > f (f (n − 1)), which contradicts the minimality of f (1). Thus, f (1) is the
smallest among {f (n) : n ≥ 1} and f (1) < f (n) for every n > 1.
Next, consider S2 = {f (n) : n ≥ 2}. Again this set has a smallest element, say f (a2 ).
Clearly, f (a2 ) > f (1). Can we show that a2 = 2? If a2 > 2 then f (a2 − 1) is well-defined.
Then f (a2 ) > f (f (a2 − 1)). This will contradict the minimality of f (a2 ) once we prove that
f (a2 − 1) ≥ 2. Since a2 − 1 ≥ 2, hence f (a2 − 1)>f (1) =⇒ f (a2 − 1) ≥ f (1) + 1 ≥ 2.
The red inequality above is due to the fact that f (1) is minimal. Thus we have shown that
a2 is indeed equal to 2, i.e., the minimal element in S2 is f (2).

We shall show the following statements simultaneously by induction on k.


(i) The minimum element in Sk = {f (n) : n ≥ k} is f (k),
(ii) f (1) < f (2) < · · · < f (k),
(iii) If n ≥ k then f (n) ≥ k.

Suppose this claim holds for all k < m. Lets prove it for k = m.
Let f (am ) be the minimal element of Sm = {f (n) : n ≥ m}. If am > m then am − 1 ≥ m, so
f (am −1) ≥ m (by hypothesis (iii)). Hence f (f (am −1)) < f (am ) contradicts the minimality
of f (am ) in Sm . Therefore we conclude that am = m, i.e., f (m) must be the minimum element
in Sm . Note that if f (m) = f (m0 ) for any m < m0 then we get a contradiction (just replace
am with m0 in the above argument).
Note that (ii) is immediate from the fact that f (m − 1) is the smallest element in Sm−1 and
hence f (m) > f (m − 1). And we already had f (m − 1) > · · · > f (2) > f (1).
Next we have to show that if n ≥ m then f (n) ≥ m. This follows from the fact that if n > m
then f (n) > f (m) and f (m) > f (m − 1) > f (m − 2) > · · · > f (1). Thus, there are at least
m numbers less than f (n), hence f (n) > m. This completes the induction.
We have shown that f (1) < f (2) < f (3) < · · · . Now, fix any n. It follows from the above
that f (n) ≥ n. Now if f (n) > n then f (n) ≥ n + 1 =⇒ f (f (n)) ≥ f (n + 1), which
contradicts the given inequality that f (f (n)) < f (n + 1). Therefore f (n) = n for all n ∈ N.

3
Problem 4. Let N0 be the set N ∪ {0}. Find all f : N0 → N0 satisfying f (m2 + n2 ) =
f (m)2 + f (n)2 for every m, n ∈ N0 and f (1) > 0.
Solution: First show that f (0) = 0 and f (m2 ) = f (m)2 and f (1) = 1. Hence we also have

f (m2 + n2 ) = f (m2 ) + f (n2 ) (2)


Now verify that f (n) = n holds small values.
f (2) = f (12 + 12 ) = f (1)2 + f (1)2 = 2
f (4) = f (22 ) = f (2)2 = 4
f (5) = f (22 + 12 ) = f (22 ) + f (12 ) = 22 + 12 = 5
f (3)2 + f (4)2 = f (32 + 42 ) = f (52 ) = f (5)2 = 32 + 42 =⇒ f (3) = 3.

Note that if p2 + q 2 = r2 + s2 then f (p)2 + f (q)2 = f (r)2 + f (s)2 . (Prove this by combining
(2) with the master equation.) So in order to show that f (n) = n for every n, by induction
on n, we can use identities of the form p2 + q 2 = r2 + s2 .

We aim to find a, b such that the term 2 · 3k · a + 2 · 4k · b equals 2 · 5k · c for some c.

(3k + a)2 + (4k + b)2 = (5k + c)2 + d

(3k + 1)2 + (4k + 3)2 = (5k + 3)2 + 12 =⇒ f (3k + 1)2 + f (4k + 3)2 = f (5k + 3)2 + f (1)2

Find such identities for f (5k), f (5k + 1), f (5k + 2), f (5k + 4) and complete the proof by
induction on k.
For a complete solution, you may consult BJV’s book (this is Problem 2.11 there).

You might also like