Prime Divisibility

For coprime positive integers n n and k k , we define ord n ( k ) \text{ord}_n (k) to be the smallest positive integer such that k ord n ( k ) 1 ( m o d n ) k^{\text{ord}_n (k)} \equiv 1 \pmod{n} .

Let L = lcm { ord 36097 ( 1 ) , ord 36097 ( 2 ) , , ord 36097 ( 36096 ) } . L= \text{lcm} \left \{ \text{ord}_{36097} (1), \text{ord}_{36097} (2) , \cdots , \text{ord}_{36097} (36096) \right \}. Find the last three digits of L + 1 L+1 .

Details and assumptions

  • To put it in words, L L is the smallest positive integer which is a multiple of all of the following numbers: ord 36097 ( 1 ) , ord 36097 ( 2 ) , , ord 36097 ( 36096 ) . \text{ord}_{36097}(1), \text{ord}_{36097} (2) , \cdots , \text{ord}_{36097} (36096) .

  • You may use the fact that 36097 36097 is a prime.

  • Even though 1 0 1 ( m o d 36097 ) 1^0 \equiv 1 \pmod{36097} , ord 36097 ( 1 ) 0 \text{ord}_{36097} (1) \neq 0 , since the order is defined to be positive. In this case, ord 36097 ( 1 ) = 1 \text{ord}_{36097} (1) = 1 .


The answer is 97.

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.

4 solutions

*Claim 1: * If k m 1 ( m o d n ) k^m \equiv 1 \pmod{n} for some m N m \in \mathbb{N} , then ord n ( k ) \text{ord}_n (k) must divide m m .

*Proof: * Write m = a × ord n ( k ) + b m = a \times \text{ord}_n (k) + b , where a , b Z + a, b \in \mathbb{Z}^+ and 0 b < ord n ( k ) 0 \leq b < \text{ord}_n (k) by the division algorithm. Note that k m 1 ( m o d n ) k a × ord n ( k ) + b 1 ( m o d n ) ( k ord n ( k ) ) a × k b 1 ( m o d n ) 1 a × k b 1 ( m o d n ) k b 1 ( m o d n ) \begin{array}{lrcl} & k^m & \equiv & 1 \pmod{n} \\ \implies & k^{a \times \text{ord}_n (k) + b} & \equiv & 1 \pmod{n} \\ \implies & \left( k^{\text{ord}_n (k)} \right) ^ a \times k^b & \equiv & 1 \pmod{n} \\ \implies & 1^a \times k^b & \equiv & 1 \pmod{n} \\ \implies & k^b & \equiv & 1 \pmod{n} \end{array} However, b > 0 b>0 contradicts the minimality of ord n ( k ) \text{ord}_n (k) . The only possible value of b b , thus, is 0 0 . It follows that ord n ( k ) \text{ord}_n (k) divides m m . \blacksquare


*Claim 2: * Let p p be any prime. For all integers n n between 2 2 and p 2 p-2 (inclusive), p 1 n + 2 n + + ( p 1 ) n . p \mid 1^n + 2^n + \cdots + (p-1)^n .

*Proof: * On the contrary, assume there exists an integer n n between 2 2 and p 2 p-2 such that p p doesn't divide 1 n + 2 n + + ( p 1 ) n 1^n + 2^n + \cdots + (p-1)^n . WLOG let n n be the smallest positive integer satisfying the property. It is well known that for all n , p N n, p \in \mathbb{N} , k = 1 n ( n + 1 k ) ( 1 k + 2 k + + p k ) = ( p + 1 ) n + 1 1. \sum \limits_{k=1}^{n} \binom{n+1}{k} \left( 1^k + 2^k + \cdots + p^k \right) = (p+1)^{n+1} - 1. So, k = 1 n 1 ( n + 1 k ) ( 1 k + 2 k + + p k ) + ( n + 1 n ) ( 1 n + 2 n + + p n ) = ( p + 1 ) n + 1 1. \sum \limits_{k=1}^{n-1} \binom{n+1}{k} \left( 1^k + 2^k + \cdots + p^k \right) + \dbinom{n+1}{n} \left( 1^n + 2^n + \cdots + p^n \right) = (p+1)^{n+1} - 1. By our assumption, p k = 1 n 1 ( n + 1 k ) ( 1 k + 2 k + + p k ) , p \left. \right | \sum \limits_{k=1}^{n-1} \binom{n+1}{k} \left( 1^k + 2^k + \cdots + p^k \right) , since n n is the smallest positive integer such that 1 n + 2 n + + p n 1^n + 2^n + \cdots + p^n is a multiple of p p . Also note that ( p + 1 ) n + 1 1 n + 1 1 0 ( m o d p ) (p+1)^{n+1} \equiv 1^{n+1} - 1 \equiv 0 \pmod{p} . It follows that p p divides ( n + 1 n ) ( 1 n + 2 n + + ( p 1 ) n ) \dbinom{n+1}{n} \left( 1^n + 2^n + \cdots + (p-1)^n \right) . But since n p 2 n \leq p-2 , gcd ( n + 1 , p ) = 1 \gcd (n+1, p) = 1 and hence p 1 n + 2 n + + ( p 1 ) n p \left. \right | 1^n + 2^n + \cdots + (p-1)^n , which is a contradiction. \blacksquare


By FLT, n 36096 1 ( m o d 36097 ) n^{36096} \equiv 1 \pmod{36097} whenever gcd ( n , 36097 ) = 1 \gcd (n, 36097) = 1 . By claim 1, it follows that ord 36097 ( k ) \text{ord}_{36097} (k) must divide 36096 36096 for all k k coprime to 36097 36097 , so L 3606 L | 3606 . Now note that 1 L + 2 L + + 3609 6 L 1 + 1 + + 1 36096 times 36096 ( m o d 36097 ) . 1^L + 2^L + \cdots + 36096^L \equiv \underbrace{1 + 1 + \cdots + 1}_{36096 \text{ times}} \equiv 36096 \pmod{36097}. By claim 2, L L cannot be less than 36096 36096 . Hence L = 36096 L= 36096 , and L + 1 = 36097 L+1= 36097 , whose last three digits are 097 \boxed{097} .

A few typos:

  • ... so L 36096 L | 36096 .

    • ...since n n is the smallest positive integer such that 1 n + 2 n + + p n 1^n + 2^n + \cdots + p^n is not a multiple of p p .

Sreejato Bhattacharya - 7 years, 4 months ago

It is detailed and best for the beginners in the modern algebra.

Moniratna Bhuyan - 7 years, 4 months ago
Ron van den Burg
Feb 10, 2014

What the problem describes is exactly Carmichael's Function λ \lambda (Reduced Totient Function) . And λ ( 36097 ) = ϕ ( 36097 ) \lambda(36097)=\phi(36097) , the Euler Totient function, because 36097 36097 is a power of an odd prime.

Euler's Totient function ϕ ( 36097 ) = 36096 \phi(36097)=36096 , so L = 36096 + 1 L=36096+1 with last three digits 097 \boxed{097} .

Indeed this all is based on Fermat's little theorem (Saurav). Then you have Euler's Totient function ϕ ( n ) \phi(n) which calculates a number that any order must divide. So, if k j 1 ( m o d n ) k^j\equiv 1 \pmod n for any k , j 0 , 1 , . . . , n 1 k, j \in {0, 1, ..., n-1} , then j j must divde ϕ ( n ) \phi(n) . Fermat's little theorem says that ϕ ( n ) = n 1 \phi(n)=n-1 if n n is prime. Euler's Totient function extends this to any positive number n.

You can find the Totient function by factoring n n in primefactors and reduce one of each prime factor: if n = p 1 a 1 p 2 a 2 p j a j n = p_1^{a_1}\cdot p_2^{a_2} \dots p_j^{a_j} , then ϕ ( n ) = ( p 1 1 ) p 1 a 1 1 ( p 2 1 ) p 2 a 2 1 ( p j 1 ) p j a j 1 \phi(n) = (p_1-1)\cdot p_1^{a_1-1}\cdot (p_2-1)\cdot p_2^{a_2-1} \dots (p_j-1)\cdot p_j^{a_j-1} .

The problem left is that the Totient function is not the smallest such a number. So, we know that the asked lcm must divide ϕ ( n ) \phi(n) , but does not need to be equal. This is where Carmichael's extension comes in: it defines the reduced totient function that is exactly the smallest number, or lcm, of all orders; the Carmichael's function is a little bit more difficult to compute. Please refer to the wikipedia page for the definition.

Furthermore, Sourav, there is no such thing as overlapping. The Carmichael's function is defined for any n n , not just for 36097.

Ron van den Burg - 7 years, 3 months ago

I never knew about Carmichael's Function! Thanks for sharing this solution. :)

Sreejato Bhattacharya - 7 years, 4 months ago

I first thought about the modulo multiplication group for 36097. But that was very much time consuming and irritating :P. Suddenly Carmichael's Function striked my mind. And it got solved so easily. :)

Moniratna Bhuyan - 7 years, 4 months ago

what if the carmichael's function had not overlapped with this number? how would we calculate then(assuming it is possible to calculate that), could you please illustrate that. Thanks

Sourav Chaudhuri - 7 years, 3 months ago

Hey, Can we use Fermat's little theorem for solving the problem? a^(p-1) is congruent to 1 mod p . Isn't Carmichael's theorem based on Fermat's work?

Saurav Zangeneh - 7 years, 4 months ago
Nick Stanford
Apr 29, 2014

From basic group theory, we know that ( Z / p ) × (\mathbb{Z}/p)^\times is a cyclic group of order p 1 p-1 for prime p p . That is, the generator of the group has order p 1 p-1 . Since the order of every element divides the order of the generator, the least common multiple of all of the orders must be p 1 p-1 .

Faraz Masroor
Feb 18, 2014

Read the hints! 36097 is prime, so by FLT anything to the 36096th power is 1 mod 36097. it is known that the order divides 36096 in this case, but it clearly can't be more than 36096 because 36096 works. We must have a primitive root by some theorem that I don't know how to prove, so 36096 is the desired LCM.

We must have a primitive root by some theorem that I don't know how to prove...

Eh? What's the name of the theorem?

Sreejato Bhattacharya - 7 years, 3 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...