具体数学--(各种基本和式)

UTF8gbsn

在《具体数学》中第二章第四节主要介绍一些基本的和形式。这一节的内容应该熟记与搞清楚。

common sums

  • ∑ j ∈ J , k ∈ K a j b k = ( ∑ j ∈ J a j ) ( ∑ k ∈ K b k ) \sum_{j\in J,k\in K}a_{j}b_{k}=(\sum_{j\in J}a_j)(\sum_{k\in K}b_k) jJ,kKajbk=(jJaj)(kKbk)

  • mapping

    ∑ j ∈ J ∑ k ∈ K ( j ) a j , k = ∑ k ∈ K ′ ∑ j ∈ J ′ ( k ) a j , k \sum_{j\in J}\sum_{k\in K(j)}a_{j,k}=\sum_{k\in K^{'}}\sum_{j\in J^{'}(k)}a_{j,k} jJkK(j)aj,k=kKjJ(k)aj,k

  • [ 1 ⩽ j ⩽ n ] [ j ⩽ k ⩽ n ] = [ 1 ⩽ j ⩽ k ⩽ n ] = [ 1 ⩽ k ⩽ n ] [ 1 ⩽ j ⩽ k ] [1\leqslant j\leqslant n][j\leqslant k\leqslant n]=[1\leqslant j\leqslant k\leqslant n]=[1\leqslant k\leqslant n][1\leqslant j\leqslant k] [1jn][jkn]=[1jkn]=[1kn][1jk]

    ∑ j = 1 n ∑ k = j n a j , k = ∑ 1 ⩽ j ⩽ k ⩽ n a j , k = ∑ k = 1 n ∑ j = 1 k a j , k \sum_{j=1}^{n}\sum_{k=j}^{n}a_{j,k}=\sum_{1\leqslant j\leqslant k\leqslant n}a_{j,k}=\sum_{k=1}^{n}\sum_{j=1}^{k}a_{j,k} j=1nk=jnaj,k=1jknaj,k=k=1nj=1kaj,k

    example 1

    [ a 1 a 1 a 1 a 2 a 1 a 3 . . . a 1 a n a 2 a 1 a 2 a 2 a 2 a 3 . . . a 2 a n a 3 a 1 a 3 a 2 a 3 a 3 . . . a 3 a n ⋮ ⋮ ⋮ ⋮ a n a 1 a n a 2 a n a 3 . . . a n a n ] \left[ \begin{array}{ccccc} a_1a_1 & a_1a_2 & a_1a_3 & ... & a_1a_n \\ a_2a_1 & a_2a_2 & a_2a_3 & ... & a_2a_n \\ a_3a_1 & a_3a_2 & a_3a_3 & ... & a_3a_n \\ \vdots & \vdots & \vdots & & \vdots \\ a_na_1 & a_na_2 & a_na_3 & ... & a_na_n \end{array} \right] a1a1a2a1a3a1ana1a1a2a2a2a3a2ana2a1a3a2a3a3a3ana3............a1ana2ana3ananan

    How about the sum of right upper triangle
    S ⊲ = ∑ 1 ⩽ j ⩽ k ⩽ n a j a k S_{\lhd}=\sum_{1\leqslant j\leqslant k\leqslant n}a_ja_k S=1jknajak

    The question here is how to simplify the formula S ⊲ S_{\lhd} S S ⊳ + S ⊲ = ∑ 1 ⩽ j , k ⩽ n a j a k + ∑ 1 ⩽ j = k ⩽ n a j a k = ( ∑ 1 ⩽ j ⩽ n a j ) 2 + ∑ i ⩽ k ⩽ n a k 2 \left. \begin{aligned} S_{\rhd}+S_{\lhd} &=\sum_{1 \leqslant j,k\leqslant n}a_ja_k+\sum_{1\leqslant j=k\leqslant n}a_ja_k\\ &=(\sum_{1\leqslant j\leqslant n}a_j)^2+\sum_{i\leqslant k \leqslant n}a_k^2 \end{aligned} \right. S+S=1j,knajak+1j=knajak=(1jnaj)2+iknak2

    from
    ∑ j = 1 n ∑ k = j n a j , k = ∑ 1 ⩽ j ⩽ k ⩽ n a j , k = ∑ k = 1 n ∑ j = 1 k a j , k \sum_{j=1}^{n}\sum_{k=j}^{n}a_{j,k}=\sum_{1\leqslant j\leqslant k\leqslant n}a_{j,k}=\sum_{k=1}^{n}\sum_{j=1}^{k}a_{j,k} j=1nk=jnaj,k=1jknaj,k=k=1nj=1kaj,k,
    we can get S ⊳ = S ⊲ S_{\rhd}=S_{\lhd} S=S

    at last

    S ⊲ = 1 2 [ ( ∑ 1 ⩽ j ⩽ n a j ) 2 + ∑ i ⩽ k ⩽ n a k 2 ] S_{\lhd} = \frac{1}{2}[(\sum_{1\leqslant j\leqslant n}a_j)^2+\sum_{i\leqslant k \leqslant n}a_k^2] S=21[(1jnaj)2+iknak2]

    example 2

    S = ∑ 1 ⩽ j < k ⩽ n ( a k − a j ) ( b k − b j ) S=\sum_{1\leqslant j <k \leqslant n}(a_k-a_j)(b_k-b_j) S=1j<kn(akaj)(bkbj)

    The question here is how to simplify the formula of S S S. Let us look
    at the fact below,

    [ 1 ⩽ j < k ⩽ n ] + [ 1 ⩽ k < j ⩽ k ] = [ 1 ⩽ j , k ⩽ n ] − [ 1 ⩽ j = k ⩽ n ] [1\leqslant j < k \leqslant n]+[1\leqslant k < j \leqslant k]=[1\leqslant j, k \leqslant n]-[1\leqslant j = k \leqslant n] [1j<kn]+[1k<jk]=[1j,kn][1j=kn]

    S = ∑ 1 ⩽ j < k ⩽ n ( a k − a j ) ( b k − b j ) = ∑ 1 ⩽ k < j ⩽ n ( a j − a k ) ( b j − b k ) S=\sum_{1\leqslant j <k \leqslant n}(a_k-a_j)(b_k-b_j)=\sum_{1\leqslant k <j \leqslant n}(a_j-a_k)(b_j-b_k) S=1j<kn(akaj)(bkbj)=1k<jn(ajak)(bjbk)

    2 S = ∑ 1 ⩽ j , k ⩽ n ( a k − a j ) ( b k − b j ) − ∑ 1 ⩽ j = k ⩽ n ( a k − a j ) ( b k − b j ) 2S=\sum_{1\leqslant j,k\leqslant n}(a_k-a_j)(b_k-b_j)-\sum_{1\leqslant j = k \leqslant n}(a_k-a_j)(b_k-b_j) 2S=1j,kn(akaj)(bkbj)1j=kn(akaj)(bkbj)

    We should aware that the second part of the left equation is zero.

    2 S = ∑ 1 ⩽ j , k ⩽ n ( a k − a j ) ( b k − b j ) 2S=\sum_{1\leqslant j,k\leqslant n}(a_k-a_j)(b_k-b_j) 2S=1j,kn(akaj)(bkbj)

    Let’s expand ( a k − a j ) ( b k − b j ) (a_k-a_j)(b_k-b_j) (akaj)(bkbj) to a k b k + a j b j − a k b j − a j b k a_kb_k+a_jb_j-a_kb_j-a_jb_k akbk+ajbjakbjajbk.
    It’s obvious that

    ∑ 1 ⩽ j , k ⩽ n a k b k = ∑ 1 ⩽ j , k ⩽ n a j b j , ∑ 1 ⩽ j , k ⩽ n a k a j = ∑ 1 ⩽ j , k ⩽ n a j a k \sum_{1\leqslant j,k\leqslant n}a_kb_k=\sum_{1\leqslant j,k\leqslant n}a_jb_j,\sum_{1\leqslant j,k\leqslant n}a_ka_j=\sum_{1\leqslant j,k\leqslant n}a_ja_k 1j,knakbk=1j,knajbj,1j,knakaj=1j,knajak

    We should get the formula below

    2 S = 2 ∑ 1 ⩽ j , k ⩽ n a k b k − 2 ∑ 1 ⩽ j , k ⩽ n a j b k 2S=2\sum_{1\leqslant j,k\leqslant n}a_kb_k-2\sum_{1\leqslant j,k\leqslant n}a_jb_k 2S=21j,knakbk21j,knajbk

    Let’s simplify the right part of the upper equation

    • ∑ 1 ⩽ j , k ⩽ n a k b k = n ∑ i ⩽ k ⩽ n a k b k \sum_{1\leqslant j,k\leqslant n}a_kb_k = n \sum_{i\leqslant k\leqslant n}a_kb_k 1j,knakbk=niknakbk

    • ∑ 1 ⩽ j , k ⩽ n a j b k = ( ∑ 1 ⩽ k ⩽ n a k ) ( ∑ 1 ⩽ k ⩽ n b k ) \sum_{1\leqslant j,k\leqslant n}a_jb_k = (\sum_{1\leqslant k\leqslant n}a_k)(\sum_{1\leqslant k\leqslant n}b_k) 1j,knajbk=(1knak)(1knbk)

    Finally,

    2 S = 2 n ∑ i ⩽ k ⩽ n a k b k − 2 ( ∑ 1 ⩽ k ⩽ n a k ) ( ∑ 1 ⩽ k ⩽ n b k ) 2S=2n \sum_{i\leqslant k\leqslant n}a_kb_k-2(\sum_{1\leqslant k\leqslant n}a_k)(\sum_{1\leqslant k\leqslant n}b_k) 2S=2niknakbk2(1knak)(1knbk)

    Rewrite the upper equation

    ( ∑ 1 ⩽ k ⩽ n a k ) ( ∑ 1 ⩽ k ⩽ n b k ) = n ∑ k = 1 n a k b k − ∑ 1 ⩽ j < k ⩽ n ( a k − a j ) ( b k − b j ) (\sum_{1\leqslant k\leqslant n}a_k)(\sum_{1\leqslant k\leqslant n}b_k) = n \sum_{k=1}^{n}a_kb_k-\sum_{1\leqslant j<k\leqslant n}(a_k-a_j)(b_k-b_j) (1knak)(1knbk)=nk=1nakbk1j<kn(akaj)(bkbj)

Chebyshev’s monotonic inequlities

We can go back to equation

( ∑ 1 ⩽ k ⩽ n a k ) ( ∑ 1 ⩽ k ⩽ n b k ) = n ∑ k = 1 n a k b k − ∑ 1 ⩽ j < k ⩽ n ( a k − a j ) ( b k − b j ) (\sum_{1\leqslant k\leqslant n}a_k)(\sum_{1\leqslant k\leqslant n}b_k) = n \sum_{k=1}^{n}a_kb_k-\sum_{1\leqslant j<k\leqslant n}(a_k-a_j)(b_k-b_j) (1knak)(1knbk)=nk=1nakbk1j<kn(akaj)(bkbj)
If a 1 ⩽ a 2 ⩽ . . . ⩽ a n a_1\leqslant a_2\leqslant ...\leqslant a_n a1a2...an and
b 1 ⩽ b 2 ⩽ . . . ⩽ b n b_1\leqslant b_2\leqslant ... \leqslant b_n b1b2...bn, we can get

( ∑ 1 ⩽ k ⩽ n a k ) ( ∑ 1 ⩽ k ⩽ n b k ) ⩽ n ∑ k = 1 n a k b k (\sum_{1\leqslant k\leqslant n}a_k)(\sum_{1\leqslant k\leqslant n}b_k)\leqslant n \sum_{k=1}^{n}a_kb_k (1knak)(1knbk)nk=1nakbk

If a 1 ⩽ a 2 ⩽ . . . ⩽ a n a_1\leqslant a_2\leqslant ...\leqslant a_n a1a2...an and
b 1 ⩾ b 2 ⩾ . . . ⩾ b n b_1\geqslant b_2\geqslant ... \geqslant b_n b1b2...bn, we can get

( ∑ 1 ⩽ k ⩽ n a k ) ( ∑ 1 ⩽ k ⩽ n b k ) ⩽ n ∑ k = 1 n a k b k (\sum_{1\leqslant k\leqslant n}a_k)(\sum_{1\leqslant k\leqslant n}b_k)\leqslant n \sum_{k=1}^{n}a_kb_k (1knak)(1knbk)nk=1nakbk

The original Chebyshev’s monotonic inequalities

[ ( ∫ a b f ( x ) d x ] [ ∫ a b g ( x ) d x ] ⩽ ( b − a ) ∫ a b f ( x ) g ( x ) d x [(\int_{a}^{b}{f(x)dx}][\int_{a}^{b}{g(x)dx}]\leqslant (b-a)\int_{a}^{b}{f(x)g(x)dx} [(abf(x)dx][abg(x)dx](ba)abf(x)g(x)dx

the condition is that f ( x ) f(x) f(x) and g ( x ) g(x) g(x) are monotonic nodecreasing
functions.

Actually

If a 1 ⩽ a 2 ⩽ . . . ⩽ a n a_1\leqslant a_2\leqslant ...\leqslant a_n a1a2...an, the
∑ k = 1 n a k b k \sum_{k=1}^{n}a_kb_k k=1nakbk will get the largest value when
b 1 ⩽ b 2 ⩽ . . ⩽ b n b_1\leqslant b_2\leqslant ..\leqslant b_n b1b2..bn, and the
∑ k = 1 n a k b k \sum_{k=1}^{n}a_kb_k k=1nakbk will get the smallest value when
b 1 ⩾ b 2 ⩾ . . . ⩾ b n b_1\geqslant b_2\geqslant ...\geqslant b_n b1b2...bn

How to proof this simple fact: (Hint1: Draw pictures, Hint2:
Recursive method)

Mapping

f : J → K f:J \rightarrow K f:JK

∑ j ∈ J a f ( j ) = ∑ k ∈ K a k # f − 1 ( k ) , f − 1 ( k ) = { j ∣ f ( j ) = k } \sum_{j\in J}a_{f(j)}=\sum_{k\in K}a_k \#f^{-1}(k), f^{-1}(k)=\{j|f(j)=k\} jJaf(j)=kKak#f1(k),f1(k)={jf(j)=k}

# \# # is how many j maps to the same k

example:

S n = ∑ 1 ⩽ j < k ⩽ n 1 k − j S_n=\sum_{1\leqslant j<k\leqslant n}\frac{1}{k-j} Sn=1j<knkj1

  • method 1 Let’s accumulate j j j first.

    S n = ∑ 1 ⩽ k ⩽ n ∑ 1 ⩽ j < k 1 k − j = ∑ 1 ⩽ k ⩽ n ∑ 1 ⩽ k − j < k 1 j = ∑ 1 ⩽ k ⩽ n ∑ 0 < j ⩽ k − 1 1 j = ∑ 1 ⩽ k ⩽ n H k − 1 = ∑ 0 ⩽ k < n H k \left. \begin{aligned} S_n&=\sum_{1\leqslant k\leqslant n}\sum_{1\leqslant j< k} \frac{1}{k-j}\\ &=\sum_{1\leqslant k\leqslant n} \sum_{1\leqslant k-j<k}\frac{1}{j}\\ &=\sum_{1\leqslant k\leqslant n} \sum_{0<j\leqslant k-1}\frac{1}{j}\\ &=\sum_{1\leqslant k\leqslant n} H_{k-1}\\ &=\sum_{0\leqslant k<n}H_k \end{aligned} \right. Sn=1kn1j<kkj1=1kn1kj<kj1=1kn0<jk1j1=1knHk1=0k<nHk

    We cannot get closed form at last.

  • method 2

    Let’s accumulate k first.

    S n = ∑ 0 ⩽ j < n H j S_n=\sum_{0\leqslant j<n}H_j Sn=0j<nHj

    It is the same as method 1

  • method 3

    Can we map k to k-j?

    S n = ∑ 1 ⩽ j < k ⩽ n 1 k − j = ∑ 1 ⩽ j < k + j ⩽ n 1 k = ∑ 1 ⩽ k ⩽ n ∑ 1 ⩽ j ⩽ n − k 1 k = ∑ 1 ⩽ k ⩽ n n − k k = ∑ 1 ⩽ k ⩽ n n k − ∑ 1 ⩽ k ⩽ n = n ∑ 1 ⩽ k ⩽ n 1 k − n = n H n − n \left. \begin{aligned} S_n&=\sum_{1\leqslant j<k\leqslant n}\frac{1}{k-j}\\ &=\sum_{1\leqslant j<k+j\leqslant n}\frac{1}{k}\\ &=\sum_{1\leqslant k\leqslant n}\sum_{1\leqslant j\leqslant n-k}\frac{1}{k}\\ &=\sum_{1\leqslant k\leqslant n}\frac{n-k}{k}\\ &=\sum_{1\leqslant k\leqslant n}\frac{n}{k}-\sum_{1\leqslant k\leqslant n}\\ &=n \sum_{1\leqslant k\leqslant n}\frac{1}{k}-n=nH_n-n \end{aligned} \right. Sn=1j<knkj1=1j<k+jnk1=1kn1jnkk1=1knknk=1knkn1kn=n1knk1n=nHnn

    Three methods have three kinds of direction of sum. (upper and
    down),(left and right) and (diagonal). ( k = 1 k = 2 k = 3 k = 4 j = 1 1 1 1 2 1 3 j = 2 1 1 1 2 j = 3 1 1 j = 4 ) \left( \begin{array}{ccccc} & k=1 & k=2 & k=3 & k=4 \\ j=1 & & \frac{1}{1} & \frac{1}{2} & \frac{1}{3} \\ j=2 & & & \frac{1}{1} & \frac{1}{2} \\ j=3 & & & & \frac{1}{1} \\ j=4 & & & & \end{array} \right) j=1j=2j=3j=4k=1k=211k=32111k=4312111

  • 0
    点赞
  • 4
    收藏
    觉得还不错? 一键收藏
  • 0
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值