求和符号

1. 裂项法：利用${\displaystyle a_{k}=b_{k+1}-b_{k}}$ 求出${\displaystyle \sum _{k=m}^{n}a_{k}}$ 。
2. 错位相减法：透过两个求和式的相减化简求和数列的求和方法。
3. 倒序求和：对于有对称中心的函数${\displaystyle f(x)+f(2a-x)=2b}$ 首尾求和^[1][2]
4. 逐项求导：可从${\displaystyle \displaystyle \sum _{k=0}^{n}x^{k}={\frac {x^{n+1}-1}{x-1}}}$ 推导出${\displaystyle \displaystyle \sum _{k=0}^{n}k^{m}x^{k}}$ ^[3]
5. 阿贝尔变换：

${\displaystyle \sum _{i=1}^{n}a_{i}b_{i}=a_{1}(b_{1}-b_{2})+(a_{1}+a_{2})(b_{2}-b_{3})+\dots +(a_{1}+a_{2}+\dots +a_{n-1})(b_{n-1}-b_{n})+(a_{1}+a_{2}+\dots +a_{n})b_{n}}$ 

以下设p为多项式，${\displaystyle \deg p(k)=m,\Delta p(k)=p(k+1)-p(k)}$ 

${\displaystyle \sum p(k)}$ 

${\displaystyle \sum p(k)}$ 是对一个多项式求和，自然数方幂和、等幂求和、等差数列求和都属于对多项式求和。

• 帕斯卡矩阵形式

  ${\displaystyle \sum _{k=1}^{n}p(k)={\begin{pmatrix}C_{n}^{1}&C_{n}^{2}&\cdots &C_{n}^{m+1}\end{pmatrix}}{\begin{pmatrix}C_{0}^{0}&0&\cdots &0\\-C_{1}^{0}&C_{1}^{1}&\cdots &0\\\vdots &\vdots &\ddots &\vdots \\(-1)^{m}C_{m}^{0}&(-1)^{m-1}C_{m}^{1}&\cdots &C_{m}^{m}\\\end{pmatrix}}{\begin{pmatrix}p(1)\\p(2)\\\vdots \\p(m+1)\end{pmatrix}}}$ ^[4]

• 差分变换形式

  ${\displaystyle p(k)=\sum _{j=1}^{m+1}C_{k-1}^{j-1}\Delta ^{j-1}p(1)}$ 

  ${\displaystyle \sum _{k=1}^{n}p(k)=\sum _{j=1}^{m+1}C_{n}^{j}\Delta ^{j-1}p(1)}$ ^[5]

${\displaystyle \sum u_{k}v_{k}x^{k}}$ 

当${\displaystyle u_{k}=p(k)}$ 为多项式，${\displaystyle \sum _{l=0}^{\infty }v_{l}x^{l}}$ 易求高阶导数时，${\displaystyle \sum _{k=0}^{\infty }u_{k}v_{k}x^{k}}$ 有封闭型和式

${\displaystyle \sum _{k=0}^{\infty }u_{k}v_{k}x^{k}=\sum _{k=0}^{\infty }{\frac {\Delta ^{k}u_{0}x^{k}}{k!}}{\frac {d^{k}}{dx^{k}}}(\sum _{l=0}^{\infty }v_{l}x^{l})}$ ^[6]

${\displaystyle \sum p(k)q^{k}}$ 

• ${\displaystyle u_{k}=p(k),v_{k}=1,x=q,\sum u_{k}v_{k}x^{k}=\sum p(k)q^{k}}$ 

  有限和${\displaystyle \displaystyle \sum _{k=1}^{n}p(k)q^{k-1}}$ 有封闭型和式

  当p为常数时，是对等比数列求和，当p为一次多项式时，是对差比数列求和。

  ${\displaystyle \displaystyle \sum _{k=1}^{n}p(k)q^{k-1}=f(n)q^{n}-f(0)}$ 

  ${\displaystyle f(n)={\frac {p(n)}{q-1}}+{\frac {1}{(q-1)^{2}}}\sum _{k=1}^{m}{\frac {(-1)^{k}q^{k-1}}{(q-1)^{k-1}}}\Delta ^{k}(p(n))={\frac {1}{q-1}}\sum _{k=0}^{m}({\frac {-q}{q-1}})^{k}\Delta ^{k}p(n+1)}$ ^[4]

${\displaystyle \sum {\frac {p(k)}{k!}}x^{k}}$ 

• ${\displaystyle u_{k}=p(k),v_{k}={\frac {1}{k!}},\sum u_{k}v_{k}x^{k}=\sum {\frac {p(k)}{k!}}x^{k}}$ 

  ${\displaystyle \sum _{n=0}^{\infty }{\frac {p(n)}{n!}}x^{n}=e^{x}\sum _{k=0}^{m}{\frac {\Delta ^{k}p(0)}{k!}}x^{k}}$ ^[7]

${\displaystyle \sum H_{k}p(k)}$ 

${\displaystyle \sum _{k=1}^{n}H_{k}p(k)=(\sum _{j=0}^{m}C_{n+1}^{j+1}\Delta ^{j}p(0))H_{n}-\sum _{j=0}^{m}{\frac {C_{n}^{j+1}}{j+1}}\Delta ^{j}p(0)}$ ，其中${\displaystyle H_{n}}$ 为调和数或调和级数

一阶求和公式

• ${\displaystyle \sum _{r=0}^{n}{\binom {n}{r}}=2^{n}}$ 
• ${\displaystyle \sum _{r=0}^{n-k}{\frac {(-1)^{r}(n+1)}{k+r+1}}{\binom {n-k}{r}}={\binom {n}{k}}^{-1}}$ 
• ${\displaystyle \sum _{r=0}^{n}{\binom {dn}{dr}}={\frac {1}{d}}\sum _{r=1}^{d}(1+e^{\frac {2\pi ri}{d}})^{dn}}$ ^{[参 1]}

• ${\displaystyle F_{n}=\sum _{i=0}^{\infty }{\binom {n-i}{i}}}$ ^{[参 2]}

${\displaystyle F_{n-1}+F_{n}=\sum _{i=0}^{\infty }{\binom {n-1-i}{i}}+\sum _{i=0}^{\infty }{\binom {n-i}{i}}=1+\sum _{i=1}^{\infty }{\binom {n-i}{i-1}}+\sum _{i=1}^{\infty }{\binom {n-i}{i}}=1+\sum _{i=1}^{\infty }{\binom {n+1-i}{i}}=\sum _{i=0}^{\infty }{\binom {n+1-i}{i}}=F_{n+1}}$ 

• ${\displaystyle \sum _{i=m}^{n}{\binom {i}{a}}={\binom {n+1}{a+1}}-{\binom {m}{a+1}}}$ 

${\displaystyle {\binom {m}{a+1}}+{\binom {m}{a}}+{\binom {m+1}{a}}...+{\binom {n}{a}}={\binom {n+1}{a+1}}}$ 

${\displaystyle \sum _{i=m}^{n}{\binom {k_{1}+i}{k_{2}}}={\binom {k_{1}+n+1}{k_{2}+1}}-{\binom {k_{1}+m}{k_{2}+1}}}$ 

${\displaystyle \sum _{i=m}^{n}{\binom {k_{1}+i}{k_{2}+i}}={\binom {k_{1}+n+1}{k_{2}+n}}-{\binom {k_{1}+m}{k_{2}+m-1}}}$ 

二阶求和公式

• ${\displaystyle \sum _{r=0}^{n}{\binom {n}{r}}^{2}={\binom {2n}{n}}}$ 
• ${\displaystyle \sum _{i=0}^{n}{\binom {r_{1}+n-1-i}{r_{1}-1}}{\binom {r_{2}+i-1}{r_{2}-1}}={\binom {r_{1}+r_{2}+n-1}{r_{1}+r_{2}-1}}}$ ^{[参 3]}

${\displaystyle (1-x)^{-r_{1}}(1-x)^{-r_{2}}=(1-x)^{-r_{1}-r_{2}}}$ 

${\displaystyle (1-x)^{-r_{1}}(1-x)^{-r_{2}}=(\sum _{n=0}^{\infty }{\binom {r_{1}+n-1}{r_{1}-1}}x^{n})(\sum _{n=0}^{\infty }{\binom {r_{2}+n-1}{r_{2}-1}}x^{n})=\sum _{n=0}^{\infty }(\sum _{i=0}^{n}{\binom {r_{1}+n-1-i}{r_{1}-1}}{\binom {r_{2}+i-1}{r_{2}-1}})x^{n}}$ 

${\displaystyle (1-x)^{-r_{1}-r_{2}}=\sum _{n=0}^{\infty }{\binom {r_{1}+r_{2}+n-1}{r_{1}+r_{2}-1}}x^{n}}$ 

• ${\displaystyle \sum _{i=0}^{k}{\binom {n}{i}}{\binom {m}{k-i}}={\binom {n+m}{k}}}$ 

范德蒙恒等式与超几何函数有关系：

${\displaystyle \sum _{i=0}^{k}{\binom {n}{i}}{\binom {m}{k-i}}={\frac {m!}{k!(m-k)!}}{}_{2}F_{1}(-n,-k;m-k+1;1)={\binom {n+m}{k}}}$ 

三阶求和公式

• ${\displaystyle {\binom {n+k}{k}}^{2}=\sum _{j=0}^{k}{\binom {k}{j}}^{2}{\binom {n+2k-j}{2k}}}$ 

范德蒙恒等式与广义超几何函数有关系：

${\displaystyle \sum _{j=0}^{k}{\binom {k}{j}}^{2}{\binom {n+2k-j}{2k}}={\frac {(n+2k)!}{(2k)!n!}}{}_{3}F_{2}(-k,-k,-n;1,-n-2k;1)={\binom {n+k}{k}}^{2}}$ 

当${\displaystyle f(x)}$ 在[a,b]单调递增时：

${\displaystyle f(a)+\int _{a}^{b}f(x)dx\leq \sum _{x=a}^{b}f(x)\leq f(b)+\int _{a}^{b}f(x)dx}$ 

当${\displaystyle f(x)}$ 在[a,b]单调递减时：

${\displaystyle f(b)+\int _{a}^{b}f(x)dx\leq \sum _{x=a}^{b}f(x)\leq f(a)+\int _{a}^{b}f(x)dx}$ ^[8]

以${\displaystyle \sum _{i=1}^{n}i^{9}}$ 为例：

• Matlab

syms k n;symsum(k^9,k,1,n)

• Mathematica

 In[1]:= Sum[i^9, {i, 1, n}]
 Out[1]:= ${\displaystyle {\frac {1}{20}}n^{2}(n+1)^{2}\left(n^{2}+n-1\right)\left(2n^{4}+4n^{3}-n^{2}-3n+1\right)}$