The Inverse Function Theorem provides sufficient conditions for a [function](/page/Function) to be locally invertible and describes the [derivative](/page/Derivative) of that inverse. It is a cornerstone of differential geometry and multivariable calculus, bridging linear approximations with local non-linear behavior. ## Motivation In single-variable calculus, a function is invertible near a point if its derivative is non-zero. In higher dimensions, we seek a similar criterion to determine when a system of non-linear equations can be solved uniquely for input variables in terms of output variables. This theorem reduces that complex non-linear problem to a linear algebraic one: checking the invertibility of the [linear map](/page/Linear%20Map) (the derivative) at that point. ## Formal Statement [quotetheorem:51] ## Examples [example:Polar Coordinates] Consider the mapping from polar to Cartesian coordinates, $f(r, \theta) = (r \cos \theta, r \sin \theta)$, defined on $U = \{(r, \theta) : r > 0\}$. The determinant of the derivative matrix at a point $(r, \theta)$ is: \begin{align*} \det(D_{(r,\theta)} f) = \det \begin{bmatrix} \cos \theta & -r \sin \theta \\ \sin \theta & r \cos \theta \end{bmatrix} = r(\cos^2 \theta + \sin^2 \theta) = r \end{align*} Since $r > 0$ on $U$, the linear map $D_{(r,\theta)} f$ is invertible everywhere. Therefore, $f$ is locally invertible at every point in $U$. [/example] [example:Quadratic Map Failure] Consider $f: \mathbb{R} \to \mathbb{R}$ defined by $f(x) = x^2$. The derivative at $x$ is the linear map $h \mapsto 2x \cdot h$. At $a = 0$, $D_0 f$ is the zero map ($h \mapsto 0$), which is not invertible. The function is not locally invertible at $0$ because every neighborhood of $0$ maps to $[0, \epsilon)$, violating injectivity. [/example] ## Consequences [theorem: Implicit Function Theorem] The [Implicit Function Theorem](/page/Implicit%20Function%20Theorem) follows directly. Let $F: \mathbb{R}^{n+m} \to \mathbb{R}^m$ be a $C^1$ map such that $F(x_0, y_0) = 0$. If the partial derivative with respect to $y$, $D_y F(x_0, y_0)$, is invertible, there exists a neighborhood of $x_0$ and a unique function $g$ such that $F(x, g(x)) = 0$. [/theorem] [proof] Construct a map $H: \mathbb{R}^{n+m} \to \mathbb{R}^{n+m}$ defined by $H(x, y) = (x, F(x, y))$. The total derivative $D_{(x_0, y_0)} H$ is represented by the block matrix: \begin{align*} [D_{(x_0,y_0)} H] = \begin{bmatrix} I_n & 0 \\ D_x F & D_y F \end{bmatrix} \end{align*} Since $D_y F$ is invertible, $\det(D H) = \det(D_y F) \neq 0$. By [theorem:InverseFunctionTheorem], $H$ has a local inverse. The components of this inverse provide the function $g$. [/proof] ## Problems [problem] Determine the [set](/page/Set) of points in $\mathbb{R}^2$ where the function $f(x, y) = (e^x \cos y, e^x \sin y)$ is locally invertible. [/problem] [solution] We compute the matrix of the derivative $D_{(x,y)} f$: \begin{align*} [D_{(x,y)} f] = \begin{bmatrix} e^x \cos y & -e^x \sin y \\ e^x \sin y & e^x \cos y \end{bmatrix} \end{align*} Next, we calculate the determinant: \begin{align*} \det([D_{(x,y)} f]) = (e^x \cos y)(e^x \cos y) - (-e^x \sin y)(e^x \sin y) = e^{2x} \end{align*} Since $e^{2x} > 0$ for all real $x$, the linear map is always invertible. Thus, $f$ is locally invertible at **all** points $(x, y) \in \mathbb{R}^2$. [/solution] [problem] Let $f: \mathbb{R}^2 \to \mathbb{R}^2$ be given by $f(x, y) = (x^2 - y^2, 2xy)$. Find the points where the Inverse Function Theorem fails to apply. [/problem] [solution] We calculate the Jacobian determinant: \begin{align*} [D_{(x,y)} f] = \begin{bmatrix} 2x & -2y \\ 2y & 2x \end{bmatrix} \implies \det([D_{(x,y)} f]) = 4x^2 + 4y^2 \end{align*} The theorem fails to apply where the derivative is not invertible (determinant is 0). \begin{align*} 4(x^2 + y^2) = 0 \iff x = 0 \text{ and } y = 0 \end{align*} Thus, the theorem applies everywhere except at the origin $(0,0)$. [/solution] ## Applications The Inverse Function Theorem is essential in: 1. **[Integration](/page/Integral) by Substitution:** The change of variables formula requires the transformation to be a diffeomorphism, verified by the determinant of $D_a f$. 2. **Differential Geometry:** It is used to prove that a graph of a function is a manifold and to define coordinate charts. 3. **Optimization:** In the method of Lagrange multipliers, it ensures the constraints define a valid surface. ## References * Rudin, *Principles of Mathematical Analysis* (1976) * Spivak, *Calculus on Manifolds* (1965) * Munkres, *Analysis on Manifolds* (1991)