多元向量值函数的微分

定理：

设：
$m, n \in \mathbb N, m, n \ge 1,$
${\mathbf{x}} = (x_1, \cdots, x_n)^\intercal,$
$f( {\mathbf{x}}) = (f_1( {\mathbf{x}}), \cdots, f_m( {\mathbf{x}}))^\intercal,$
$f'( {\mathbf{x}}) = J_f( {\mathbf{x}})= \mathbf{A} = \left ( a_{ij} ( {\mathbf{x}}) \right )_{m \times n} = (a_1, \cdots, a_n)^\intercal,$
$\Delta {\mathbf{x}} = (\Delta x_1, \cdots, \Delta x_n)^\intercal,$
$\Delta {\mathbf{y}} = f( {\mathbf{x}} + \Delta {\mathbf{x}}) - f( {\mathbf{x}}) = (\Delta y_1, \cdots, \Delta y_m)^\intercal,$
r=∑nj=1Δxj<