Chain Rule
Let
\[ \mathbb{R}^n \xrightarrow{f} \mathbb{R}^m \xrightarrow{g} \mathbb{R}^k \]
be differentiable functions. Then the composition \(g\circ f\) is differentiable and
\[ (g\circ f)'(x_1,\dots,x_n) = g'(f(x_1,\dots,x_n))f'(x_1,\dots,x_n). \]
The product on the right hand side of the equation is of a \(k\times m\) matrix and an \(m\times n\) matrix, which produces a \(k\times n\) matrix, which is the size of the derivative \((g\circ f)'(x_1,\dots,x_n)\).
To shorten notation, we often load the variables \(x_1,\dots,x_n\) into a vector \(\mathbf{x}\) and write the chain rule as \[ (g\circ f)'(\mathbf{x}) = g'(f(\mathbf{x}))f'(\mathbf{x}). \]
Consider the functions \[ \mathbb{R}^3 \xrightarrow{f} \mathbb{R} \xrightarrow{g} \mathbb{R}^2 \] where \[ f(x,y,z) = x^2z\sin{y}, \quad g(u) = (u^2, 3u^3). \] Compute the derivative \((g\circ f)'(x,y,z)\) using the chain rule.
Consider the functions \[ \mathbb{R} \xrightarrow{h} \mathbb{R}^3 \xrightarrow{k} \mathbb{R}^2 \] where \[ h(t) = (e^t, t, 2t^2-t), \quad k(\alpha,\beta,\gamma) = (\alpha^2 + \beta^2, \gamma^3). \] Compute the derivative \((k\circ h)'(t)\) using the chain rule.
Chain Rule
Let
\[ \mathbb{R}^n \xrightarrow{f} \mathbb{R}^m \xrightarrow{g} \mathbb{R}^k \]
be differentiable functions. Then the composition \(g\circ f\) is differentiable and
\[ (g\circ f)'(\mathbf{x}) = g'(f(\mathbf{x}))f'(\mathbf{x}). \]
The \((i,j)\)-th entry of the derivative \((g\circ f)'(\mathbf{x})\) is the partial derivative \[ \frac{\partial (g\circ f)_i}{\partial x_j}(\mathbf{x}). \]
If you compute the \((i,j)\)-th entry of the product \(g'(f(\mathbf{x}))f'(\mathbf{x})\), you will find that it is equal to \[ \sum_{p=1}^m \frac{\partial g_i}{\partial u_p}(f(\mathbf{x}))\frac{\partial f_p}{\partial x_j}(\mathbf{x}), \] where \(u_1,\dots,u_m\) are the variables in the codomain of \(f\).
Chain Rule (partial derivatives)
Let
\[ \mathbb{R}^n \xrightarrow{f} \mathbb{R}^m \xrightarrow{g} \mathbb{R}^k \]
be differentiable functions. Then, for each \(i=1,\dots,k\) and \(j=1,\dots,n\), we have \[ \frac{\partial (g\circ f)_i}{\partial x_j}(\mathbf{x}) = \sum_{p=1}^m \frac{\partial g_i}{\partial u_p}(f(\mathbf{x}))\frac{\partial f_p}{\partial x_j}(\mathbf{x}), \] where \(u_1,\dots,u_m\) are the variables in the domain of \(g\).
Let \(g(u,v) = u^2v - v^2\), where \(u = y\sin{x}\) and \(v = xe^y\). Compute \(\displaystyle\frac{\partial g}{\partial x}\) and \(\displaystyle\frac{\partial g}{\partial y}\) using the chain rule.
Let \(h(x,y) = x\sin{x} + y^2\), where \(x = t^2 + 1\) and \(y = e^t\). Compute \(\displaystyle\frac{dh}{dt}\) using the chain rule.
Let \(\theta(r,s) = (r^2 + s^2, rs)\), where \(r = \alpha\beta\) and \(s = \alpha^2 - \beta^2\). Compute all four partial derivatives \[ \frac{\partial \theta_1}{\partial \alpha}, \quad \frac{\partial \theta_1}{\partial \beta}, \quad \frac{\partial \theta_2}{\partial \alpha}, \quad \frac{\partial \theta_2}{\partial \beta} \] using the chain rule.
The radius of a right circular cylinder is increasing at a rate of \(6\) inches per minute, and the height is decreasing at a rate of \(4\) inches per minute. What are the rates of change of the volume and surface area of the cylinder when the radius is \(12\) inches and the height is \(36\) inches?