Sixfold fun in 2015, Part 38

S = d 12090 ϕ ( d ) d S=\sum_{d|12090}\frac{\phi(d)}{d}

Find the smallest positive integer a a such that a × S a\times S is an integer.

Clarifications

  • The sum S S is taken over all 32 positive integer divisors d d of 12090 = 6 × 2015 12090=6\times 2015

  • ϕ ( d ) \phi(d) denotes the Euler's totient function .


The answer is 806.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

3 solutions

Aareyan Manzoor
Dec 21, 2015

it is a multiplicative function. let f ( d ) = ϕ ( d ) d f(d)=\dfrac{\phi(d)}{d} then f(1)=1, and f(ab) for coprime a,b f ( a b ) = ϕ ( a b ) a b = ϕ ( a ) a ϕ ( b ) b = f ( a ) f ( b ) f(ab)=\dfrac{\phi(ab)}{ab}=\dfrac{\phi(a)}{a}\dfrac{\phi(b)}{b}=f(a)f(b) the summation at 1 is one. that is also multiplicative. so d 2 3 5 13 31 ϕ ( d ) d = ( d 2 ϕ ( d ) d ) ( d 3 ϕ ( d ) d ) ( d 5 ϕ ( d ) d ) ( d 13 ϕ ( d ) d ) ( d 31 ϕ ( d ) d ) \sum_{d|2*3*5*13*31}\dfrac{\phi(d)}{d}=\left(\sum_{d|2}\dfrac{\phi(d)}{d}\right)\left(\sum_{d|3}\dfrac{\phi(d)}{d}\right)\left(\sum_{d|5}\dfrac{\phi(d)}{d}\right)\left(\sum_{d|13}\dfrac{\phi(d)}{d}\right)\left(\sum_{d|31}\dfrac{\phi(d)}{d}\right) = ( 1 + 1 2 ) ( 1 + 2 3 ) ( 1 + 4 5 ) ( 1 + 12 13 ) ( 1 + 30 31 ) = 13725 806 =\left(1+\dfrac{1}{2}\right)\left(1+\dfrac{2}{3}\right)\left(1+\dfrac{4}{5}\right)\left(1+\dfrac{12}{13}\right)\left(1+\dfrac{30}{31}\right)=\dfrac{13725}{\Large{\boxed{806}}}

Yes, very nicely done (+1)!

Otto Bretscher - 5 years, 5 months ago
Prasun Biswas
Dec 23, 2015

Very nice problem! My approach was quite different from the posted solutions. For the sake of variety, let me add my solution here too.

We note n : = 12090 = 2 × 3 × 5 × 13 × 31 n:=12090=2\times 3\times 5\times 13\times 31 .

S : = d n ϕ ( d ) d = d n d P ( d ) d = d n P ( d ) S:=\sum_{d\mid n}\frac{\phi(d)}{d}=\sum_{d\mid n}\frac{d\cdot\mathcal{P}(d)}{d}=\sum_{d\mid n}\mathcal{P}(d)

Here, P ( d ) = i ( 1 1 q ( d , i ) ) \large\mathcal{P}(d)=\prod\limits_i\left(1-\frac 1{q(d,i)}\right) where the q ( d , i ) q(d,i) 's are the prime factors of d d , i.e., we have { q ( d , i ) } { 2 , 3 , 5 , 13 , 31 } = { p i } d : d n \{q(d,i)\}\subseteq\{2,3,5,13,31\}=\{p_i\}~\forall~d:d\mid n

Since multiplicity of every prime factor of n n is 1 1 , it makes our work simple. We calculate r i = ( 1 1 p i ) r_i=\large\left(1-\frac 1{p_i}\right) for every element of { p i } i 1 = { 2 , 3 , 5 , 13 , 31 } \{p_i\}_{i\geq 1}=\{2,3,5,13,31\} .

Note that,

S = d n P ( d ) = 1 + k = 1 5 1 i 1 < i 2 < < i k 5 j = 1 k r i j \large S=\sum_{d\mid n}\mathcal{P}(d)=1+\sum_{k=1}^5\sum\limits_{1\leq i_1 < i_2 < \cdots < i_k \leq 5}\prod_{j=1}^k r_{i_j}

Now, consider the polynomial f ( x ) = ( x 1 ) i = 1 5 ( x r i ) = x 6 + a 5 x 5 + a 4 x 4 + a 1 x + a 0 (say) f(x)=(x-1)\prod\limits_{i=1}^5\left(x-r_i\right)=x^6+a_5x^5+a_4x^4+\cdots a_1x+a_0~\textrm{(say)}

By Vieta's formulas and with a bit of inspection, we can conclude that f ( 1 ) = 1 + k = 0 5 ( 1 ) k a k = 2 S f(-1)=1+\sum\limits_{k=0}^5(-1)^ka_k=2S

So, we calculate f ( 1 ) f(-1) which is,

2 S = f ( 1 ) = ( 1 1 ) i = 1 5 ( 1 r i ) = ( 1 1 ) ( 1 1 2 ) ( 1 2 3 ) ( 1 4 5 ) ( 1 12 13 ) ( 1 30 31 ) = 13725 403 \begin{aligned}2S=f(-1)&=\left(-1-1\right)\prod_{i=1}^5\left(-1-r_i\right)\\&=\left(-1-1\right)\left(-1-\frac 12\right)\left(-1-\frac 23\right)\left(-1-\frac 45\right)\left(-1-\frac{12}{13}\right)\left(-1-\frac{30}{31}\right)=\frac{13725}{403}\end{aligned}

S = 13725 806 \implies S=\frac{13725}{806}

Since gcd ( 13725 , 806 ) = 1 \gcd(13725,806)=1 , for a S = 13725 a 806 aS=\dfrac{13725a}{806} to be an integer, we need a 0 ( m o d 806 ) a\equiv 0\pmod{806} , i.e., a = 806 m m Z a=806m~\forall~m\in\Bbb Z with the smallest positive integer a a being 806 806 at m = 1 m=1 .

Moderator note:

Good approach. however it is not clear to me what you chose to define the polynomial, as opposed to making the multiplicative observation directly.

Otto Bretscher
Dec 21, 2015

Here is a compressed version of Aareyan's fine solution:

f ( n ) = d n ϕ ( d ) d f(n)=\sum_{d|n}\frac{\phi(d)}{d} is multiplicative, with f ( p ) = ϕ ( p ) p + ϕ ( 1 ) 1 = 2 p 1 p f(p)=\frac{\phi(p)}{p}+\frac{\phi(1)}{1}=\frac{2p-1}{p} for primes p p , so that f ( 2 3 5 13 31 ) = f ( 2 ) f ( 3 ) f ( 5 ) f ( 13 ) f ( 31 ) = 3 5 9 25 61 2 3 5 13 31 = 13725 806 f(2*3*5*13*31)=f(2)f(3)f(5)f(13)f(31)=\frac{3*5*9*25*61}{2*3*5*13*31}=\frac{13725}{806} .

As an interesting aside, note that f ( n ) = 1 n k = 1 n gcd ( k , n ) f(n)=\frac{1}{n}\sum_{k=1}^{n}\gcd(k,n) .

yes: f ( n ) = d n n d ϕ ( d ) n = 1 n d n n d ϕ ( d ) = 1 n d n d ϕ ( n d ) f(n)=\sum_{d|n} \dfrac{\dfrac{n}{d}\phi(d)}{n}=\dfrac{1}{n}\sum_{d|n} \dfrac{n}{d}\phi(d)=\dfrac{1}{n}\sum_{d|n} d\phi(\dfrac{n}{d}) the part after 1 n \frac{1}{n} is the formula for the sum of gcd. this is a great prove of that bieng multiplicative as the original sum f was easily proven to be multiplicative. multiplying by n doesnt change multiplicity.

Aareyan Manzoor - 5 years, 5 months ago

Log in to reply

Yes, exactly! Small typo: The first two " ϕ ( n ) " "\phi(n)" in your formula for f ( n ) f(n) should be " ϕ ( d ) " "\phi(d)" .

It does follow directly from the definition of a multiplicative function that k = 1 n gcd ( k , n ) = n d n ϕ ( d ) d \sum_{k=1}^{n}\gcd(k,n)=n\sum_{d|n}\frac{\phi(d)}{d} is multiplicative. It is trivial to verify that products and quotients of multiplicative functions are multiplicative and that the divisor sum F ( n ) = d n f ( d ) F(n)=\sum_{d|n}f(d) of a multiplicative function f ( d ) f(d) is multiplicative. (A slightly fancier explanation uses the Dirichlet convolution.)

Otto Bretscher - 5 years, 5 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...