[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

1282.0. "Enumeration Quickie: Count the boundary components of hypercube" by NOEDGE::HERMAN (Franklin B. Herman DTN 291-0170 PDM1-1/J9) Wed Aug 08 1990 21:22

    Quickie:

	Provide and validate a simple expression for 
	N(r,n), the number r-dimensional boundary 
	hypercubes of a given n-dimensional hypercube,
	e.g.,
	    
	    N(2,3) = 6  is the number of faces of a cube.

    -Franklin
T.RTitleUserPersonal
Name
DateLines
1282.1as Andrew would say, "let there be a graph"ALLVAX::JROTHIt's a bush recording...Thu Aug 09 1990 10:0825
�	Provide and validate a simple expression for 
�	N(r,n), the number r-dimensional boundary 
�	hypercubes of a given n-dimensional hypercube,

    N(r,n) = 2^(n-r) * (n choose r)

    square:

	2^1 * (2 choose 1) = 4 sides
	2^2 * (2 choose 0) = 4 vertices
	
    cube:

	2^1 * (3 choose 2) = 6 faces
	2^2 * (3 choose 1) = 12 edges
	2^3 * (3 choose 0) = 8 vertices

    tesseract:

	2^1 * (4 choose 3) = 8 3-faces
	2^2 * (4 choose 2) = 24 2-faces
	2^3 * (4 choose 1) = 32 1-faces (edges)
	2^4 * (4 choose 0) = 16 0-faces (vertices)

    - Jim
1282.2An explanation , pleaseNOEDGE::HERMANFranklin B. Herman DTN 291-0170 PDM1-1/J9Thu Aug 09 1990 10:4710
Re: .1

>    N(r,n) = 2^(n-r) * (n choose r)

    Ok ... but would you please explain your reasoning. 

    Thanks in advance,

    -Franklin
1282.3validation...ALLVAX::JROTHIt's a bush recording...Thu Aug 09 1990 13:335
    Any of the (n choose r) r-faces incident on one of the vertices
    can be translated along any subset of the (n-r) coordinates
    in the complement of the space spanned by the face.

     - Jim
1282.4Euler Characteristic of S^nNOEDGE::HERMANFranklin B. Herman DTN 291-0170 PDM1-1/J9Thu Aug 09 1990 20:1295
    Re: .3

	Jim, I really appreciated your direct, elegant and concise 
    argument (especially in light of the way I attacked it!). It took
    me an embarrassing long time to finally grok the part about 
    "translated along along any subset of the (n-r) coordinates in 
    the complement" but it was well worth the effort.

	For the benefit of any of the other readers who may struggle 
    with combinatorial arguments as I do, and did with this particular 
    one along with me, I understood it to mean that by an translating 
    an r-face through  a choice of subset the (n-r) coordinates 
    is the same choosing whether 
    or not to translate (either forward or backward depending on which 
    keeps the face inside the cube) in each of the (n-r) complementary 
    directions. E.g., for a suitably choosen edge (1-face) of a cube 
    there are four choices of translations: 

	    (00) - stay put
	    (10) - move across
	    (01) - move down
	    (11) - move across and down (order doesn't matter)
	    
	I came up a with reductionist approach which is far less direct 
    but perhaps interesting. It was motivated by wanting to compute the 
    Euler Characteristic of the boundary, Bd(I^n), of the n-dimenensional 
    hypercube, I^n (and therefore the Euler Characteristic of S^(n-1) the
    (n-1)-dimensional hyper-sphere -- its continuous equivalent. I'm
    aware that this can also be done with simplexes as is standard in
    a course in Alg. Top., however I thought a cube approach might lead
    to a geometrically simplier argument.)

	Reductive observation:
    
	The n-dimenensional hypercube, I^n, can be generated by
    unit parallel translation of an I^(n-1). In the process, 
    each of the I^(n-2) boundary component faces of the 
    moving I^(n-1), sweeps out a unique I^(n-1) face of I^n. 
    If we add to this, the final and initial position of the 
    moving I^(n-1) face, we obtain the relation: 
	
	N(n-1,n) = N(n-2,n-1) + 2

	More generally, each I^(r-1) component of the sweeping
    I^(n-1) face generates a unique I^r component of I^n. This together
    with the initial and final positions of each I^r component in the	
    the sweeping I^(n-1) yields the recurrance relation:

	N(r,n) = N(r-1,n-1) + 2 N(r,n-1)

    With the aim of eventually computing the Euler characteristic
    (a generalization of the V-F+E invariant for surfaces), consider
    the polynomial C_n(x) in x of degree n with coefficients N(r,n):

                  n               
                 ___         r    
        C_n(x) = >   N(r,n) x     
                 ---              
                 r=0              

    Using the above recurrance relation we obtain 

	C_n(x) = x C_n-1(x) + 2 C_n-1(x) = (2 + x) C_n-1(x)

    Which inductively results in 

                        n       
	C_n(x) = (2 + x)  , whose coefficients yield 

                             n-r
	    N(r,n) = c(r,n) 2    , where c(r,n) is "r choose n".
           
    Since the Euler characteristic is just the alternating sum of the 
    number of boundary components of fixed dimension, we get

                                  n                
                      n          ___            r     
        Euler_char (I)      =    >   N(r,n) (-1)   = C_n(-1) = 1
                                 ---               
                                 r=0               

    Finally, the Euler characteristic of Bd(I^n) is just the C_n(-1) minus
    its last term , we get:

                                               n            n
	Euler_char (Bd(I^n))  =  C_n(-1) - (-1) =   1 - (-1)     


		       /  0 if n is even 
		    = |
		       \  2 if n is odd


    -Franklin