亚导数
- 当函数不可微时,不可计算出其普通的导数,此时便需要引入亚导数
- Example:
函数 y = ∣ x ∣ y=|x| y=∣x∣ 不可微,其亚导数为
∂ ∣ x ∣ ∂ x = { 1 , x > 0 − 1 , x < 0 a , x = 0 , a ∈ [ 0 , 1 ] \frac{\partial |x|}{\partial x}=\begin{equation} \left\{ \begin{array}{lr} 1, x>0 & \\ -1,x<0 & \\ a, x=0, a\in [0,1] & \end{array} \right. \end{equation} ∂x∂∣x∣=⎩ ⎨ ⎧1,x>0−1,x<0a,x=0,a∈[0,1]
将导数拓展到向量
- y是标量,x是标量,导数也是标量;
- y是标量,x是向量,导数是向量;
- y是向量,x是标量,导数是向量;
- y是向量,x也是向量,导数是一个矩阵。
- Example1:
当 y y y是标量, x ⃗ = [ x 1 , x 2 , . . . , x n ] T \vec{x}=[x_1,x_2,...,x_n]^T x=[x1,x2,...,xn]T为向量时,有 ∂ y ∂ x ⃗ = [ ∂ y ∂ x 1 , ∂ y ∂ x 2 , . . . , ∂ y ∂ x n ] \frac{\partial y}{\partial \vec{x}}=[\frac{\partial y}{\partial x_1},\frac{\partial y}{\partial x_2},...,\frac{\partial y}{\partial x_n}] ∂x∂y=[∂x1∂y,∂x2∂y,...,∂xn∂y]
Explain:- < u , v > <u,v> <u,v>表示向量的内积, < u , v > ′ = u T ∂ v ∂ x + v T ∂ u ∂ x <u,v>'=u^T\frac{\partial v}{\partial x}+v^T\frac{\partial u}{\partial x} <u,v>′=uT∂x∂v+vT∂x∂u公式简略推导如下:
- Example2:
当 y ⃗ = [ y 1 , y 2 , . . . , y n ] T \vec{y}=[y_1,y_2,...,y_n]^T y=[y1,y2,...,yn]T是向量, x x x是标量时,有 ∂ y ⃗ ∂ x = [ ∂ y 1 ∂ x , ∂ y 2 ∂ x , . . . , ∂ y n ∂ x ] T \frac{\partial \vec{y}}{\partial x}=[\frac{\partial y_1}{\partial x},\frac{\partial y_2}{\partial x},...,\frac{\partial y_n}{\partial x}]^T ∂x∂y=[∂x∂y1,∂x∂y2,...,∂x∂yn]T - Example3:
当 y ⃗ = [ y 1 , y 2 , . . . , y n ] T \vec{y}=[y_1,y_2,...,y_n]^T y=[y1,y2,...,yn]T是向量, x ⃗ = [ x 1 , x 2 , . . . , x n ] T \vec{x}=[x_1,x_2,...,x_n]^T x=[x1,x2,...,xn]T也是向量时有