更广义的莱布尼兹公式

“Product Rule” for Differential Operators

• 首先上一个传统的莱布尼茨公式！

$$(uv{)}^{(n)}=\sum _{k=0}^n{C}_n^k{u}^{(n-k)}{v}^{(k)}$$

 

• to prove the following identity:
• L be a linear differential operator of order n
  • characteristic polynomial $p_n$.
• Then for functions $v,y$

 

• 例如$L=a{D}^2+bD+c$

$L[vy]=a(v{y}^{\prime \prime }+2v^{\prime }{y}^{\prime }+v^{\prime \prime }y)+b(v^{\prime }y+v{y}^{\prime })+cvy$

$=v(a{y}^{\prime \prime }+b{y}^{\prime }+cy)+v^{\prime }(2a{y}^{\prime }+by)+\frac{1}{2}{v}^{\prime \prime }(2ay)$

• 咋证明呢？

ode和matlab结合的东西

$y^{(4)}-5y^{(3)}+7y^{(2)}-3y^{\prime }=2e^t(4t^2+2t+7)$

• 似乎$y=t^2(\frac{-t^2}{3}+\frac{t}{3}-6)e^t$是他的结果
• 下面验证一下结果！

syms t;
y = t^2*(-1/3*t^2 + 1/3 *t -6)*exp(t);
result = diff(y,t,4) - 5* diff(y,t,3) + 7 *diff(y,t,2)-3 * diff(y,t,1);
simplify(result)

 

• 首先$(r-1{)}^{2}r(r-3)$
• 所以有一个特解长这样子
  • $t^2(c_2{t}^2+c_{1}t+c_0)e^t$

所以 L [ { t 2 ( c 2 t 2 + c 1 t + c 0 ) } e t ] L[\{t^2(c_2t^2 + c_1 t + c_0)\}e^t] L[{t2(c2​t2+c1​t+c0​)}et]

$ = $

k	$\frac{1}{k!}$	$v^{(k)}$	$p^{(k)}$	$p^{(k)}(1)$
0	1	$c_2{t}^4+c_1{t}^3+c_0{t}^2$	$r^4-5r^3+7r^2-3r$	0
1	1	$4c_2{t}^3+3c_1{t}^2+2c_{0}t$	$4r^3-15r^2+14r-3$	0
2	$\frac{1}{2}$	$12c_2{t}^2+6c_{1}t+2c_0$	$12r^2-30r+14$	-4
3	$\frac{1}{6}$	$24c_{2}t+6c_1$	$24r-30$	-6
4	$\frac{1}{24}$	$24c_2$	$24$	24

• 所以

$$-2(12c_2{t}^2+6c_{1}t+2c_0)-(24c_{2}t+6c_1)+24c_2=2(4t^2+2t+7)$$

$$-24c_2=8$$
$$-12c_1-24c_2=4$$
$$-4c_0-6c_1+24c_2=14$$

• $c_0=-6$
• $c_1=\frac{1}{3}$
• $c_2=\frac{-1}{3}$

二维求和交换

$$L[vy]=\sum _{i=0}^n{a}_i(vy{)}^{(i)}$$

$=\sum _{i=0}^n{a}_i\sum _{k=0}^i{C}_i^k{v}^{(k)}{y}^{(i-k)}$

• 以上是二维求和，调换维度变成下面这样了！

$=\sum _{k=0}^n{v}^{(k)}\sum _{i=k}^n{C}_i^k{a}_i{y}^{(i-k)}$

$=\sum _{k=0}^n{v}^{(k)}\sum _{i=k}^n\frac{(i)(i-1)...(i-k+1)}{k!}{a}_i{y}^{(i-k)}$
$=\sum _{k=0}^n\frac{v^{(k)}}{k!}\sum _{i=k}^n(i)(i-1)...(i-k+1)a_i{y}^{(i-k)}$
$=\sum _{k=0}^n\frac{v^{(k)}}{k!}{L}^{(k)}[y]$

 

• $L^{(k)}$到底长啥样子呢？

$L=\sum _{i=0}^n{a}_i{x}^i$

$L^{(k)}=\sum _{i=k}^n(i)(i-1)...(i-k+1)a_i{x}^{i-k}$