Difference between revisions of "2012 USAMO Problems/Problem 4"
(→Solution) |
(→Solution: cleaned up) |
||
| Line 4: | Line 4: | ||
==Solution== | ==Solution== | ||
| − | + | By the first condition we have <math>f(1)=f(1!)=f(1)!</math> and <math>f(2)=f(2!)=f(2)!</math>, so <math>f(1)=1</math> or <math>2</math> and similarly for <math>f(2)</math>. By the second condition, we have | |
<cmath> | <cmath> | ||
| − | n\cdot n!=(n+1)!-n! | + | n\cdot n!=(n+1)!-n! \mid f(n+1)!-f(n)! \qquad \qquad (1) |
</cmath> | </cmath> | ||
| − | (1) | + | for all positive integers <math>n</math>. |
| + | |||
| + | Suppose that for some <math>n \geq 2</math> we have <math>f(n) = 1</math>. We claim that <math>f(k)=1</math> for all <math>k\ge n</math>. Indeed, from Equation (1) we have <math>f(n+1)!\equiv 1 \mod n\cdot n!</math>, and this is only possible if <math>f(n+1)=1</math>; the claim follows by induction. | ||
| − | |||
We now divide into cases: | We now divide into cases: | ||
| Line 21: | Line 22: | ||
This implies <math>f(n)=1</math> for all <math>n\ge 2</math>, but this does not satisfy the initial conditions. Indeed, we would have | This implies <math>f(n)=1</math> for all <math>n\ge 2</math>, but this does not satisfy the initial conditions. Indeed, we would have | ||
<cmath> | <cmath> | ||
| − | 3-1 | + | 3-1 \mid f(3)-f(1) |
</cmath> | </cmath> | ||
| − | so <math>2 | + | and so <math>2\mid -1</math>, a contradiction. |
'''Case 3:''' <math>f(1)=1</math>, <math>f(2)=2</math> | '''Case 3:''' <math>f(1)=1</math>, <math>f(2)=2</math> | ||
| − | We claim <math>f(n)=n</math> always by induction. The | + | We claim <math>f(n)=n</math> always by induction. The base cases are <math>n = 1</math> and <math>n = 2</math>. Fix <math>k > 1</math> and suppose that <math>f(k)=k</math>. By Equation (1) we have that |
<cmath> | <cmath> | ||
| − | f(k+1)\equiv k! \mod k\cdot k! | + | f(k+1)! \equiv k! \mod k\cdot k! . |
</cmath> | </cmath> | ||
| − | + | This implies <math>f(k+1)<2k</math> (otherwise <math>f(k+1)!\equiv 0 \mod k\cdot k!</math>). Also we have | |
<cmath> | <cmath> | ||
| − | k+1-1 | + | (k+1)-1 \mid f(k+1)-f(1) |
</cmath> | </cmath> | ||
| − | so <math>f(k+1)\equiv 1 \mod k</math>. This gives the solutions <math>f(k+1)=1</math> (obviously impossible | + | so <math>f(k+1)\equiv 1 \mod k</math>. This gives the solutions <math>f(k+1)=1</math> and <math>f(k+1)=k+1</math>. The first case is obviously impossible, so <math>f(k + 1) = k + 1</math>, as desired. By induction, <math>f(n) = n</math> for all <math>n</math>. This also satisfies the requirements. |
'''Case 4:''' <math>f(1)=f(2)=2</math> | '''Case 4:''' <math>f(1)=f(2)=2</math> | ||
| Line 43: | Line 44: | ||
f(k+1)\equiv 2 \mod k\cdot k! | f(k+1)\equiv 2 \mod k\cdot k! | ||
</cmath> | </cmath> | ||
| − | + | and so <math>f(k+1)<2k</math>. Also note that | |
<cmath> | <cmath> | ||
| − | k+1-1 | + | k+1-1 \mid f(k+1)-2 |
</cmath> | </cmath> | ||
| + | and | ||
<cmath> | <cmath> | ||
| − | k+1-2 | + | k+1-2 \mid f(k+1)-2 |
</cmath> | </cmath> | ||
| − | so <math>f(k+1)\equiv 2 \mod k(k-1)</math>. Then the only possible solution is <math>f(k+1)=2</math>. By induction | + | so <math>f(k+1)\equiv 2 \mod k(k-1)</math>. Then the only possible solution is <math>f(k+1)=2</math>. By induction, <math>f(n) = 2</math> for all <math>n</math>, and this satisfies all requirements. |
| + | |||
| − | + | In summary, there are three solutions: <math>\boxed{f(n)=1, f(n)=2, f(n)=n}</math>. | |
==See also== | ==See also== | ||
Revision as of 10:10, 21 June 2012
Problem
Find all functions
(where
is the set of positive integers) such that
for all positive integers
and such that
divides
for all distinct positive integers
,
.
Solution
By the first condition we have
and
, so
or
and similarly for
. By the second condition, we have
for all positive integers
.
Suppose that for some
we have
. We claim that
for all
. Indeed, from Equation (1) we have
, and this is only possible if
; the claim follows by induction.
We now divide into cases:
Case 1:
This gives
always from the previous claim, which is a solution.
Case 2:
This implies
for all
, but this does not satisfy the initial conditions. Indeed, we would have
and so
, a contradiction.
Case 3:
,
We claim
always by induction. The base cases are
and
. Fix
and suppose that
. By Equation (1) we have that
This implies
(otherwise
). Also we have
so
. This gives the solutions
and
. The first case is obviously impossible, so
, as desired. By induction,
for all
. This also satisfies the requirements.
Case 4:
We claim
by a similar induction. Again if
, then by (1) we have
and so
. Also note that
and
so
. Then the only possible solution is
. By induction,
for all
, and this satisfies all requirements.
In summary, there are three solutions:
.
See also
| 2012 USAMO (Problems • Resources) | ||
| Preceded by Problem 3 |
Followed by Problem 5 | |
| 1 • 2 • 3 • 4 • 5 • 6 | ||
| All USAMO Problems and Solutions | ||