Serving the Quantitative Finance Community

 
User avatar
TrueTears
Topic Author
Posts: 0
Joined: July 16th, 2011, 9:26 am

Divisibility question.

July 19th, 2011, 8:15 am

Hi everybody! I am new to this forum and have just joined recently. The forum seems great and I've always wanted to find and join an online finance forum Anyways, I am currently an undergraduate student in Pure Mathematics, Finance and Actuarial studies, so here's a little brainteaser combinatorics question Do there exist at least 10,000 10 digit numbers divisible by 7, all of which can be obtained from one another by a reordering of the digits?
Last edited by TrueTears on July 18th, 2011, 10:00 pm, edited 1 time in total.
 
User avatar
MCarreira
Posts: 64
Joined: January 1st, 1970, 12:00 am

Divisibility question.

July 19th, 2011, 2:37 pm

The answer (counting) is yes, no clever answer (divisibility) yet, but it probably involves the number of solutions of the equation resulting from the repeated application of the divisibility rule to a 10-digit number, where the variables are the digits.
 
User avatar
TrueTears
Topic Author
Posts: 0
Joined: July 16th, 2011, 9:26 am

Divisibility question.

July 19th, 2011, 4:37 pm

Correct, can you think of a purely combinatorial approach? That would be quite slick ;P
 
User avatar
MCarreira
Posts: 64
Joined: January 1st, 1970, 12:00 am

Divisibility question.

July 19th, 2011, 9:45 pm

For the permutations of {0,1,2,...,9} with exactly k elements, there seems to be always a particular choice that produces (k-1)! numbers that are divisible by 7 (respecting no zero on the left):k=3: {1,4,7}k=4: (1,3,5,8}k=5: {1,2,7,8,9}It will take a while to check this for k>5, but it looks interesting.
 
User avatar
TrueTears
Topic Author
Posts: 0
Joined: July 16th, 2011, 9:26 am

Divisibility question.

July 20th, 2011, 7:01 am

Haha the good ol' brute force style
 
User avatar
cm27874
Posts: 2
Joined: July 2nd, 2007, 12:10 pm

Divisibility question.

July 26th, 2011, 7:11 am

The number of 10-digit numbers modulo reordering of digits (i.e., the number of orbits of the action of the symmetric group S_10 by permutation of digits) is equal to the number of monomials of degree 10 in 10 variables, i.e. C(19,10) = 92378.Therefore, by the pigeon hole principle, there must be an orbit containing at least10^10 / ( 7 * 93278 ) > 15464 numbers divisible by 7.
 
User avatar
DevonFangs
Posts: 0
Joined: November 9th, 2009, 1:49 pm

Divisibility question.

July 26th, 2011, 12:30 pm

QuoteOriginally posted by: cm27874The number of 10-digit numbers modulo reordering of digits (i.e., the number of orbits of the action of the symmetric group S_10 by permutation of digits) is equal to the number of monomials of degree 10 in 10 variables, i.e. C(19,10) = 92378.Therefore, by the pigeon hole principle, there must be an orbit containing at least10^10 / ( 7 * 93278 ) > 15464 numbers divisible by 7.WTF, dude.
 
User avatar
pk14
Posts: 0
Joined: February 15th, 2006, 12:43 am

Divisibility question.

July 28th, 2011, 2:38 pm

Consider all the integers generated by reordering the 9 digit number 123456789. Let's denote this set by S. Clearly, |S| = 9!.So, there must exist k in {0, 1, 2, 3, 4, 5, 6} such that S_k = { x in S such that x mod 7 = k } has more than 9!/7 elements. In other words, |S_k| = 9!/7 > 50000.All the 9!/7 numbers in the set 10*S_k mod 7 = 10*k. All numbers in 10*S_k has 10 digits.Find the number a in {0, 1, 2, 3, 4, 5, 6} such that a + 10*k mod 7 = 0.Now, all numbers in 10*S_k + a are divisible by 7 and are of 10 digits and are generated by reordering 123456789a.