84 - The Multivariable Chain Rule
The chain rule is different when it comes to multivariable calculus.
Suppose that we want the derivative with respect to some parameter
Then, this is defined as
A much more intuitive form of representation of the MV Chain rule is recognizing that the above sum can be interpreted as a dot product, and that's exactly it.
We define the multivariable chain rule in vector form as $$\begin{aligned}\displaystyle
&\frac{d}{dt}f(x(t),y(t)) = \pdv{f}{x}\cdot \dfrac{dx}{dt}+ \pdv{f}{y} \cdot \dfrac{dy}{dt}=\left[\begin{matrix}\displaystyle\pdv{f}{x} \ \displaystyle\pdv{f}{y} \ \vdots\end{matrix}\right]
\cdot
\left[\begin{matrix} \dfrac{dx}{dt} \ \dfrac{dy}{dt} \ \vdots \end{matrix}\right]
\\
&= \nabla f(\vec{v}(t)) \cdot \vec{v}'(t);, \text{which can be expressed in terms of}
\
&\text{the directional derivative as } \nabla_{\large\vec{v}'(t)}f(\vec{v}(t))
\end{aligned}