Exercise 1.4.8 (Dominoes theorem)

Give three proofs that

m = 0 M ( m + k k ) = ( k + M + 1 k + 1 ) .

(a)
With k fixed, induct on M , using Theorem 1.20.
(b)
Let 𝒮 = { 1 , 2 , , k + M + 1 } . Count the number of subsets 𝒜 of 𝒮 containing k + 1 elements, with the maximum one being k + m + 1 .
(c)
Compute the coefficient of z M in the identity ( 1 + z + z 2 + ) 1 ( 1 z ) k + 1 = 1 ( 1 z ) k + 2 .

Answers

Proof.

(a)
To avoid the induction, using Pascal’s relation (1.14), ( k + m k ) + ( k + m k + 1 ) = ( k + m + 1 k + 1 ) ,

we obtain

m = 0 M ( k + m k ) = m = 0 M [ ( k + m + 1 k + 1 ) ( k + m k ) ] = m = 0 M ( k + m + 1 k + 1 ) m = 0 M ( k + m k + 1 ) = n = 1 M + 1 ( k + n k + 1 ) m = 0 M ( k + m k + 1 ) ( n = m + 1 ) = m = 1 M + 1 ( k + m k + 1 ) m = 0 M ( k + m k + 1 ) = ( k + M + 1 k + 1 ) + m = 1 M ( k + m k + 1 ) m = 1 M ( k + m k + 1 ) ( k k + 1 ) = ( k + M + 1 k + 1 )

since ( k k + 1 ) = 0 . This gives the first proof of

m = 0 M ( m + k k ) = ( k + M + 1 k + 1 ) .

(b)
Here 𝒮 = [ [ 1 , k + M + 1 ] ] . For every m [ [ 0 , M ] ] , consider the set 𝒞 m of subsets 𝒜 of 𝒮 containing k + 1 elements, with the maximum one being k + m = 1 : 𝒞 m = { 𝒜 𝒫 k + 1 ( 𝒮 ) max ( 𝒜 ) = k + m + 1 } .

Every 𝒜 𝒫 k + 1 ( 𝒮 ) has a maximum m , where k + 1 m k + M + 1 , thus m = m + k + 1 , where 0 m M . This shows that

𝒫 k + 1 ( 𝒮 ) = m = 0 M 𝒞 m .

Moreover this union is disjoint: if 𝒜 𝒞 m 𝒞 n , then m = max ( 𝒜 ) = n , thus

m n 𝒞 m 𝒞 n = .

Therefore

| 𝒫 k + 1 ( 𝒮 ) | = m = 0 M | 𝒞 m | . (1)

Moreover | 𝒫 k + 1 ( 𝒮 ) | = ( k + M + 1 k + 1 ) .

Consider the two maps

φ { 𝒫 k ( [ [ 1 , k + m ] ] ) 𝒞 m { k + m + 1 } ψ { 𝒞 m 𝒫 k ( [ [ 1 , m ] ] ) 𝒜 𝒜 { max ( 𝒜 ) }

This makes sense: if 𝒫 k ( [ [ 1 , k + m ] ] ) , then max ( { k + m + 1 } ) = k + m + 1 , where | { k + m + 1 } ) | = k + 1 , so φ ( ) 𝒞 m , and if 𝒜 𝒞 m , then max ( A ) = k + m + 1 , thus 𝒜 { max ( 𝒜 ) [ [ 1 , k + m ] ] , where | 𝒜 { max ( 𝒜 ) | = k , thus ψ ( 𝒜 ) 𝒫 k ( [ [ 1 , m ] ] ) .

Moreover, for all 𝒫 k ( [ [ 1 , k + m ] ] ) ,

( ψ ϕ ) ( ) = ( { k + m + 1 } ) { k + m + 1 } = ,

and for all 𝒜 𝒞 m ,

( φ ψ ) ( 𝒜 ) = ( 𝒜 { k + m + 1 ) } ) { k + m + 1 } = 𝒜 .

Therefore ψ φ and φ ψ are identities, thus φ is bijective, so

| 𝒞 m | = | 𝒫 k ( [ [ 1 , k + m ] ] ) | = ( k + m k ) ,

thus (1) gives

( k + M + 1 k + 1 ) = m = 0 M ( k + m k ) .

This is the second proof.

(c)
Starting from 1 ( 1 z ) k + 2 = 1 1 z 1 ( 1 z ) k + 1 ,

we obtain, using the exercise 1.4.7,

M = 0 ( k + 1 + M M ) z M = n = 0 z n m = 0 ( k + m m ) z m .

The absolute convergence of these series for | z | < 1 (*) gives the Cauchy’s product

n = 0 z n m = 0 ( k + m m ) z m = M = 0 n + m = M ( k + m m ) z M = M = 0 m = 0 M ( k + m m ) z M .

The equality of the coefficient of z M in these two expansions gives

( k + 1 + M M ) = m = 0 M ( k + m m ) .

This is the third proof.

(*) We can avoid this argument if we use asymptotic expansions instead of series expansions (or we can use formal series).

User profile picture
2024-07-26 09:49
Comments