[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference rusure::math

Title:Mathematics at DEC
Moderator:RUSURE::EDP
Created:Mon Feb 03 1986
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2083
Total number of notes:14613

258.0. "Combinatorial Problem" by HARE::STAN () Thu Apr 11 1985 17:06

An acquaintance of mine recently had the following interesting
problem published in the Monthly:

If c and m are positive integers each greater than 1, find the

number n(c,m) of ordered c-tuples (n , n , ... n ) with entries
				    1   2       c

from the initial segment {1,2,...,m} of the positive integers

such that n  < n   and n  <= n  <= ... <= n  .
	   2    1       2     3            c

			Reference
			---------
Dennis Spellman, Problem E3086, American mathematical Monthly,
		 92(Apr 1985)287.
T.RTitleUserPersonal
Name
DateLines
258.1LATOUR::JMUNZERThu Apr 25 1985 12:3630
I believe I have the answer.

[1]  Original problem:  n(c, m), where

          n  > n  <= n  <= ... <= n  .
	   1    2     3            c

Answer:  n (c, m) = m * C (c + m - 2, c - 1) - C (c + m - 1, c),

	where C (i, j) = number of ways to take a subset of j things from
	a set of i things, i. e.

		C (i, j) = i! / j! / (i - j)!

Approach:  solve easier problems below

[2]  o(c, m), where

          n  <  n  <  n  <  ... <  n  .
	   1     2     3            c

[3]  p(c, m), where

          n  <= n  <= n  <= ... <= n  .
	   1     2     3            c

[4]  (n + p)(c, m) -- essentially n  is unrestricted.
                                   1

Loved the problem.  Thank you.
258.2LATOUR::JMUNZERFri Apr 26 1985 13:5998
		FULLER EXPLANATION OF 258.1


[0]	All functions below are talking about (ordered) c-tuples

		(n1, n2, n3, ..., nc)

	where each of the n's is in { 1, 2, 3, ..., m }

[1]	Original problem:

	n (c, m) = number of c-tuples such that

		n1 > n2 <= n3 <= n4 <= ... <= nc

	E.g.  n (3, 3) = 8, because there are { 211, 212, 213, 311, 312, 313,

	322, 323 }.

	General answer will follow in [6].

[2]	Easier problem:

	o (c, m) = number of c-tuples such that

		n1 < n2 < n3 < n4 < ... < nc

	These c-tuples correspond 1-1 to the subsets of c things taken from
	a set of m distinct things, or

		C (m, c)

	which is my notation for 

		m! / c! / (m - c)!

[3]	Another easier problem:

	p (c, m) = number of c-tuples such that

		n1 <= n2 <= n3 <= n4 <= ... <= nc

	These non-strictly-increasing c-tuples correspond 1-1 to the

	strictly-increasing c-tuples

		(n1 + 0) < (n2 + 1) < (n3 + 2) < (n4 + 3) < ... < (nc + c - 1)

	which are taken from a larger set of values, namely:

		{ 1, 2, 3, ..., m + c - 1 }

	Thus,

		p (c, m) = o (c, m + c - 1) = C (m + c - 1, c)

[4]	Another easier problem:

	q (c, m) = number of c-tuples such that

		n1 is unrestricted (i. e. 1 <= n1 <= m), and

		n2 <= n3 <= n4 <= ... <= nc

	Since there are m choices for n1,

		q (c, m) = m * p (c - 1, m) = m * C (m + c - 2, c - 1)

[5]	Claim:  n + p = q

	Given n2 <= n3 <= n4 <= ... <= nc, either

		n1 > n2		[that's n]

	or

		n1 <= n2	[that's p]

[6]	Answer to original problem:

		n (c, m) = q (c, m) - p (c, m) =

			m * C (m + c - 2, c - 1) - C (m + c - 1, c)

	E. g.  n (3, 3) = 3 * C (4, 2) - C (5, 3)

			= 3 * (4! / 2! / 2!) - (5! / 3! / 2!)

			= 18 - 10

	Sets:  { 211, 212, 213, 311, 312, 313, 322, 323 } =

	{ 111, 112, 113, 122, 123, 133, 211, 212, 213, 222, 223, 233,

	 	311, 312, 313, 322, 323, 333 } -

	{ 111, 112, 113, 122, 123, 133, 222, 223, 233, 333 }