2023 AMC 12A Problems/Problem 22: Difference between revisions
No edit summary |
|||
| (13 intermediate revisions by 6 users not shown) | |||
| Line 1: | Line 1: | ||
==Problem== | ==Problem== | ||
Let < | Let <imath>f</imath> be the unique function defined on the positive integers such that <cmath>\sum_{d\mid n}d\cdot f\left(\frac{n}{d}\right)=1</cmath> for all positive integers <imath>n</imath>. What is <imath>f(2023)</imath>? | ||
< | <imath>\textbf{(A)}~-1536\qquad\textbf{(B)}~96\qquad\textbf{(C)}~108\qquad\textbf{(D)}~116\qquad\textbf{(E)}~144</imath> | ||
== Solution 1 (Very Thorough) == | == Solution 1 (Very Thorough) == | ||
First, we note that < | First, we note that <imath>f(1) = 1</imath>, since the only divisor of <imath>1</imath> is itself. | ||
Then, let's look at < | Then, let's look at <imath>f(p)</imath> for <imath>p</imath> a prime. We see that | ||
<cmath>\sum_{d \mid p} d \cdot f\left(\frac{p}{d}\right) = 1</cmath> | <cmath>\sum_{d \mid p} d \cdot f\left(\frac{p}{d}\right) = 1</cmath> | ||
<cmath>1 \cdot f(p) + p \cdot f(1) = 1</cmath> | <cmath>1 \cdot f(p) + p \cdot f(1) = 1</cmath> | ||
| Line 14: | Line 14: | ||
Nice. | Nice. | ||
Now consider < | Now consider <imath>f(p^k)</imath>, for <imath>k \in \mathbb{N}</imath>. | ||
<cmath>\sum_{d \mid p^k} d \cdot f\left(\frac{p^k}{d}\right) = 1</cmath> | <cmath>\sum_{d \mid p^k} d \cdot f\left(\frac{p^k}{d}\right) = 1</cmath> | ||
<cmath>1 \cdot f(p^k) + p \cdot f(p^{k-1}) + p^2 \cdot f(p^{k-2}) + \dotsc + p^k f(1) = 1</cmath>. | <cmath>1 \cdot f(p^k) + p \cdot f(p^{k-1}) + p^2 \cdot f(p^{k-2}) + \dotsc + p^k f(1) = 1</cmath>. | ||
It can be (strongly) inductively shown that < | It can be (strongly) inductively shown that <imath>f(p^k) = f(p) = 1-p</imath>. Here's how. | ||
We already showed < | We already showed <imath>k=1</imath> works. Suppose it holds for <imath>k = n</imath>, then | ||
<cmath>1 \cdot f(p^n) + p \cdot f(p^{n-1}) + p^2 \cdot f(p^{n-2}) + \dotsc + p^n f(1) = 1 \implies f(p^m) = 1-p \; \forall \; m \leqslant n</cmath> | <cmath>1 \cdot f(p^n) + p \cdot f(p^{n-1}) + p^2 \cdot f(p^{n-2}) + \dotsc + p^n f(1) = 1 \implies f(p^m) = 1-p \; \forall \; m \leqslant n</cmath> | ||
For < | For <imath>k = n+1</imath>, we have | ||
<cmath>1 \cdot f(p^{n+1}) + p \cdot f(p^{n}) + p^2 \cdot f(p^{n-1}) + \dotsc + p^{n+1} f(1) = 1</cmath>, then using < | <cmath>1 \cdot f(p^{n+1}) + p \cdot f(p^{n}) + p^2 \cdot f(p^{n-1}) + \dotsc + p^{n+1} f(1) = 1</cmath>, then using <imath>f(p^m) = 1-p \; \forall \; m \leqslant n</imath>, we simplify to | ||
<cmath>1 \cdot f(p^{n+1}) + p \cdot (1-p) + p^2 \cdot (1-p) + \dotsc + p^n \cdot (1-p) + p^{n+1} f(1) = 1</cmath> | <cmath>1 \cdot f(p^{n+1}) + p \cdot (1-p) + p^2 \cdot (1-p) + \dotsc + p^n \cdot (1-p) + p^{n+1} f(1) = 1</cmath> | ||
| Line 34: | Line 34: | ||
<cmath>f(p^{n+1}) + p = 1 \implies f(p^{n+1}) = 1-p</cmath>. | <cmath>f(p^{n+1}) + p = 1 \implies f(p^{n+1}) = 1-p</cmath>. | ||
Very nice! Now, we need to show that this function is multiplicative, i.e. < | Very nice! Now, we need to show that this function is multiplicative, i.e. <imath>f(pq) = f(p) \cdot f(q)</imath> for <imath>\textbf{distinct}</imath> <imath>p,q</imath> prime. | ||
It's pretty standard, let's go through it quickly. | It's pretty standard, let's go through it quickly. | ||
<cmath>\sum_{d \mid pq} d \cdot f\left(\frac{pq}{d}\right) = 1</cmath> | <cmath>\sum_{d \mid pq} d \cdot f\left(\frac{pq}{d}\right) = 1</cmath> | ||
| Line 42: | Line 42: | ||
Great! We're almost done now. | Great! We're almost done now. | ||
Let's actually plug in < | Let's actually plug in <imath>2023 = 7 \cdot 17^2</imath> into the original formula. | ||
<cmath>\sum_{d \mid 2023} d \cdot f\left(\frac{2023}{d}\right) = 1</cmath> | <cmath>\sum_{d \mid 2023} d \cdot f\left(\frac{2023}{d}\right) = 1</cmath> | ||
<cmath>1 \cdot f(2023) + 7 \cdot f(17^2) + 17 \cdot f(7 \cdot 17) + 7 \cdot 17 \cdot f(17) + 17^2 \cdot f(7) + 7 \cdot 17^2 \cdot f(1) = 1</cmath> | <cmath>1 \cdot f(2023) + 7 \cdot f(17^2) + 17 \cdot f(7 \cdot 17) + 7 \cdot 17 \cdot f(17) + 17^2 \cdot f(7) + 7 \cdot 17^2 \cdot f(1) = 1</cmath> | ||
| Line 53: | Line 53: | ||
So plugging ALL that in, we have | So plugging ALL that in, we have | ||
<cmath>f(2023) = 1 - \left(7 \cdot (-16) + 17 \cdot (-6) \cdot (-16) + 7 \cdot 17 \cdot (-16) + 17^2 \cdot (-6) + 7 \cdot 17^2\right)</cmath> | <cmath>f(2023) = 1 - \left(7 \cdot (-16) + 17 \cdot (-6) \cdot (-16) + 7 \cdot 17 \cdot (-16) + 17^2 \cdot (-6) + 7 \cdot 17^2\right)</cmath> | ||
~ < | which, be my guest simplifying, is <imath>\boxed{\textbf{(B)} \ 96}</imath> | ||
~ <imath>\color{magenta} zoomanTV</imath> | |||
== Solution 2 == | == Solution 2 == | ||
| Line 63: | Line 64: | ||
So now we get | So now we get | ||
<cmath>\frac{1}{n}= \sum_{d\mid n}\frac{f(d)}{d}</cmath> | <cmath>\frac{1}{n}= \sum_{d\mid n}\frac{f(d)}{d}</cmath> | ||
Also, notice that both < | Also, notice that both <imath>\frac{f(d)}{d}</imath> and <imath>\frac{1}{n}</imath> are arithmetic functions. Applying Möbius inversion formula, we get | ||
<cmath>\frac{f(n)}{n}=\sum_{d\mid n}\frac{ \mu (d) }{\frac{n}{d} }=\frac{1}{n} \sum_{d\mid n}d\cdot \mu (d) </cmath> | <cmath>\frac{f(n)}{n}=\sum_{d\mid n}\frac{ \mu (d) }{\frac{n}{d} }=\frac{1}{n} \sum_{d\mid n}d\cdot \mu (d) </cmath> | ||
So | So | ||
<cmath>f(n)=1-p_1-p_2-...+p_1p_2+...=(1-p_1)(1-p_2)...=\prod_{p\mid n}(1-p)</cmath> | <cmath>f(n)=1-p_1-p_2-...+p_1p_2+...=(1-p_1)(1-p_2)...=\prod_{p\mid n}(1-p)</cmath> | ||
So the answer should be < | So the answer should be <imath>f(2023)=\prod_{p\mid 2023}(1-p)=(1-7)(1-17)=\boxed{\textbf{(B)} \ 96}</imath> | ||
~ZZZIIIVVV | ~ZZZIIIVVV | ||
== Solution 3 == | |||
From the problem, we want to find <imath>f(2023)</imath>. Using the problem, we get <imath>f(2023)+7f(289)+17f(119)+119f(17)+289f(7)+2023f(1)=1</imath>. By plugging in factors of <imath>2023</imath>, we get | |||
<cmath> | |||
\begin{align} | |||
f(7)+7f(1)=1\\ | |||
f(17)+17f(1)=1\\ | |||
f(119)+7f(17)+17f(7)+119f(1)=1\\ | |||
f(289)+17f(17)+289f(1)=1 | |||
\end{align} | |||
</cmath> | |||
Notice that <imath>(4)-17(2)=f(289)</imath>, so <imath>f(289)=-16</imath>. Similarly, notice that <imath>(3)-17(1)=f(119)+7f(17)=-16</imath>. Now, substituting this all back into our equation to solve for <imath>f(2023)</imath>, we get | |||
<cmath> | |||
\begin{align*} | |||
f(2023)+7f(289)+17(f(119)+7f(17))+289(f(7)+7f(1))=1\\ | |||
f(2023)+7 \cdot (-16) + 17 \cdot (-16) + 289 \cdot (1) = 1\\ | |||
f(2023)=\boxed{\textbf{(B)} \ 96} | |||
\end{align*} | |||
</cmath> | |||
-PhunsukhWangdu | |||
==Solution 4== | |||
Consider any <imath>n \in \Bbb N</imath> with prime factorization <imath>n = \Pi_{i=1}^k p_i^{\alpha_i}</imath>. | |||
Thus, the equation given in this problem can be equivalently written as | |||
<cmath> | |||
\[ | |||
\sum_{\beta_1 = 0}^{\alpha_1} | |||
\sum_{\beta_2 = 0}^{\alpha_2} | |||
\cdots | |||
\sum_{\beta_k = 0}^{\alpha_k} | |||
\Pi_{i=1}^k p_i^{\alpha_i - \beta_i} | |||
\cdot | |||
f \left( \Pi_{i=1}^k p_i^{\beta_i} \right) | |||
= 1 . | |||
\] | |||
</cmath> | |||
<imath>\noindent \textbf{Special case 1}</imath>: <imath>n = 1</imath>. | |||
We have <imath>f \left( 1 \right) = 1</imath>. | |||
<imath>\noindent \textbf{Special case 2}</imath>: <imath>n</imath> is a prime. | |||
We have | |||
<cmath> | |||
\[ | |||
1 \cdot f \left( n \right) + n \cdot f \left( 1 \right) = 1 . | |||
\] | |||
</cmath> | |||
Thus, <imath>f \left( n \right) = 1 - n</imath>. | |||
<imath>\noindent \textbf{Special case 3}</imath>: <imath>n</imath> is the square of a prime, <imath>n = p_1^2</imath>. | |||
We have | |||
<cmath> | |||
\[ | |||
1 \cdot f \left( p_1^2 \right) + p_1 \cdot f \left( p_1 \right) + p_1^2 \cdot f \left( 1 \right) = 1. | |||
\] | |||
</cmath> | |||
Thus, <imath>f \left( p_1^2 \right) = 1 - p_1</imath>. | |||
<imath>\noindent \textbf{Special case 4}</imath>: <imath>n</imath> is the product of two distinct primes, <imath>n = p_1 p_2</imath>. | |||
We have | |||
<cmath> | |||
\[ | |||
1 \cdot f \left( p_1 p_2 \right) + p_1 \cdot f \left( p_2 \right) + p_2 \cdot f \left( p_1 \right) + p_1 p_2 \cdot f \left( 1 \right) = 1. | |||
\] | |||
</cmath> | |||
Thus, <imath>f \left( p_1 p_2 \right) = 1 - p_1 - p_2 + p_1 p_2</imath>. | |||
<imath>\noindent \textbf{Special case 5}</imath>: <imath>n</imath> takes the form <imath>n = p_1^2 p_2</imath>, where <imath>p_1</imath> and <imath>p_2</imath> are two distinct primes. | |||
We have | |||
<cmath> | |||
\[ | |||
1 \cdot f \left( p_1^2 p_2 \right) + p_1 \cdot f \left( p_1 p_2 \right) + p_1^2 \cdot f \left( p_2 \right) + p_2 \cdot f \left( p_1^2 \right) | |||
+ p_1 p_2 f \left( p_1 \right) + p_1^2 p_2 f \left( 1 \right) = 1. | |||
\] | |||
</cmath> | |||
Thus, <imath>f \left( p_1^2 p_2 \right) = 1 - p_1 - p_2 + p_1 p_2</imath>. | |||
The prime factorization of 2023 is <imath>7 \cdot 17^2</imath>. | |||
Therefore, | |||
<cmath> | |||
\begin{align*} | |||
f \left( 2023 \right) & = 1 - 7 - 17 + 7 \cdot 17 \\ | |||
& = \boxed{\textbf{(B) 96}}. | |||
\end{align*} | |||
</cmath> | |||
~Steven Chen (Professor Chen Education Palace, www.professorchenedu.com) | |||
==Solution 5 (Very much overkill, but fast if you know stuff)== | |||
In the below, denote <imath>\operatorname{id}(x)=x</imath> and <imath>1(x)=1</imath>, and let the Dirichlet convolution of <imath>f</imath> and <imath>g</imath> be <imath>f*g</imath>. Denote the Dirichlet inverse of <imath>f</imath> as <imath>f^{-1}</imath>. Also, let the Dirichlet series of <imath>f</imath> evaluated at <imath>s</imath> be: | |||
<cmath>DG_f(s)=\sum_{n=1}^{\infty}\frac{f(n)}{n^s}.</cmath> | |||
Let <imath>\zeta(s)</imath> be the Riemann zeta function. | |||
We note that <imath>\sum_{d|n}{d\cdot f\left(\frac{n}{d}\right)}=1</imath> is equivalent to <imath>f=1*\operatorname{id}^{-1}</imath>. We state a well-known theorem: | |||
<cmath>DG_f(s)DG_g(s)=DG_{f*g}(s)</cmath>. | |||
By this: | |||
<cmath>DG_{\operatorname{id}^{-1}}(s)=\frac{1}{DG_{\operatorname{id}}(s)}=\frac{1}{\zeta(s-1)}.</cmath> | |||
We know also that: | |||
<cmath>\sum_{n=1}^{\infty}\frac{\mu(n)}{n^s}=\frac{1}{\zeta(s)}</cmath>, | |||
so: | |||
<cmath>DG_{\operatorname{id}^{-1}}(s)=\sum_{n=1}^{\infty}\frac{n\mu(n)}{n^s},</cmath> | |||
i.e., <imath>\operatorname{id}^{-1}(n)=n\mu(n).</imath> Thus: | |||
\begin{align*} | |||
f(2023)&=\left(\operatorname{id}^{-1}*1\right)(2023)\\ | |||
&=\sum_{d|2023}{n\mu(n)}\\ | |||
&=\mu(1)+7\mu(7)+17\mu(17)+7\cdot 17\mu(7\cdot 17)\\ | |||
&= 7\cdot 17-7-17+1\\ | |||
&=96 | |||
\end{align*} | |||
Hence we choose <imath>\boxed{(B)}</imath>. | |||
~Sivin | |||
==Video Solution by MOP 2024== | ==Video Solution by MOP 2024== | ||
| Line 79: | Line 203: | ||
https://youtu.be/Trz8DEmgAtk | https://youtu.be/Trz8DEmgAtk | ||
==Video Solution== | |||
https://youtu.be/Fyd1hGGHZ8k | |||
~Steven Chen (Professor Chen Education Palace, www.professorchenedu.com) | |||
==See also== | ==See also== | ||
Latest revision as of 09:20, 11 November 2025
Problem
Let
be the unique function defined on the positive integers such that
for all positive integers
. What is
?
Solution 1 (Very Thorough)
First, we note that
, since the only divisor of
is itself.
Then, let's look at
for
a prime. We see that
Nice.
Now consider
, for
.
.
It can be (strongly) inductively shown that
. Here's how.
We already showed
works. Suppose it holds for
, then
For
, we have
, then using
, we simplify to
.
Very nice! Now, we need to show that this function is multiplicative, i.e.
for
prime.
It's pretty standard, let's go through it quickly.
Using our formulas from earlier, we have
Great! We're almost done now.
Let's actually plug in
into the original formula.
Let's use our formulas! We know
So plugging ALL that in, we have
which, be my guest simplifying, is
~
Solution 2
First, change the problem into an easier form.
So now we get
Also, notice that both
and
are arithmetic functions. Applying Möbius inversion formula, we get
So
So the answer should be
~ZZZIIIVVV
Solution 3
From the problem, we want to find
. Using the problem, we get
. By plugging in factors of
, we get
Notice that
, so
. Similarly, notice that
. Now, substituting this all back into our equation to solve for
, we get
-PhunsukhWangdu
Solution 4
Consider any
with prime factorization
.
Thus, the equation given in this problem can be equivalently written as
:
.
We have
.
:
is a prime.
We have
Thus,
.
:
is the square of a prime,
.
We have
Thus,
.
:
is the product of two distinct primes,
.
We have
Thus,
.
:
takes the form
, where
and
are two distinct primes.
We have
Thus,
.
The prime factorization of 2023 is
.
Therefore,
~Steven Chen (Professor Chen Education Palace, www.professorchenedu.com)
Solution 5 (Very much overkill, but fast if you know stuff)
In the below, denote
and
, and let the Dirichlet convolution of
and
be
. Denote the Dirichlet inverse of
as
. Also, let the Dirichlet series of
evaluated at
be:
Let
be the Riemann zeta function.
We note that
is equivalent to
. We state a well-known theorem:
.
By this:
We know also that:
,
so:
i.e.,
Thus:
\begin{align*}
f(2023)&=\left(\operatorname{id}^{-1}*1\right)(2023)\\
&=\sum_{d|2023}{n\mu(n)}\\
&=\mu(1)+7\mu(7)+17\mu(17)+7\cdot 17\mu(7\cdot 17)\\
&= 7\cdot 17-7-17+1\\
&=96
\end{align*}
Hence we choose
.
~Sivin
Video Solution by MOP 2024
https://YouTube.com/watch?v=gdhVqdRhMsQ
~r00tsOfUnity
Video Solution by OmegaLearn
Video Solution
~Steven Chen (Professor Chen Education Palace, www.professorchenedu.com)
See also
| 2023 AMC 12A (Problems • Answer Key • Resources) | |
| Preceded by Problem 21 |
Followed by Problem 23 |
| 1 • 2 • 3 • 4 • 5 • 6 • 7 • 8 • 9 • 10 • 11 • 12 • 13 • 14 • 15 • 16 • 17 • 18 • 19 • 20 • 21 • 22 • 23 • 24 • 25 | |
| All AMC 12 Problems and Solutions | |
These problems are copyrighted © by the Mathematical Association of America.