The Implicit Function Theorem: A Proof

-

We prove the Implicit Function Theorem for the bivariate case.

See also:



Theorem (Implicit function theorem)

Let \(F(x,y)\) be a function of class \(C^1\) on an open region \(U \subset \mathbb{R}^2\). Suppose the point \(P= (a,b)\) satisfies the following conditions:
  • \(F(a,b) = 0\) (i.e., \(P\) is a point on the curve \(F(x,y) = 0\)),
  • \(F_y(a,b) \neq 0\).
Then, there exist an open interval \(I\) on the \(x\)-axis such that \(a \in I\) and a univariate function \(y = \varphi(x)\) on \(I\) such that
  1. \(F(x,\varphi(x)) = 0\) for all \(x \in I\), that is, \(\varphi(x)\)  is an implicit function of \(F(x,y) = 0\), and
  2. \(b = \varphi(a)\).
Furthermore, the function \(\varphi(x)\) is differentiable on \(I\) and
\[\varphi'(x) = -\frac{F_x(x,\varphi(x))}{F_y(x,\varphi(x))}.\tag{Eq:IF}\]

Proof. Since \(F_y(a,b) \neq 0\), either \(F_y(a,b) > 0\) or \(F_y(a,b) < 0\). In the following, we assume \(F_y(a,b) > 0\) (the other case is similar). We prove the theorem in three steps.

Step 1. Constructing a potential implicit function \(\varphi(x)\).

Since \(F_y(a,b) > 0\) and \(F_y(x,y)\) is continuous, we can find a sufficiently small \(\varepsilon >0\) such that
\[|x - a| < \varepsilon, |y - b| < 2\varepsilon \implies F_y(x,y) > 0.\]
(That is, if \((x,y)\) is sufficiently close to \((a,b)\), then \(F_y(x,y) > 0\).) For all \(x_0\)satisfying \(|x_0 - a| < \varepsilon\), consider the univariate function \(g(y) = F(x_0, y)\). Then, on the interval \((b - 2\varepsilon, b + 2\varepsilon)\), \(g'(y) = F_y(x_0, y) > 0\) so \(g(y)\) is a strictly monotone increasing function. In particular, \(F(a,b) = 0\) implies \(F(a, b-\varepsilon) < 0\) and \(F(a, b + \varepsilon) > 0\).

Again, since \(F_y(x,y)\) is continuous, we can find a sufficiently small \(\varepsilon' > 0\) such that, for all \(x_0\) such that \(|x_0 - a| < \varepsilon'\), the following hold:
  • The univariate function \((x_0,y)\) of \(y\) is strictly monotone increasing on \((b - 2\varepsilon, b + 2\varepsilon)\),
  •  \(F(x_0, b - \varepsilon) < 0\) and \(F(x_0, b + \varepsilon) > 0\).
By the Intermediate Value Theorem, for all \(x_0\) such that \(|x_0 - a| < \varepsilon'\), there exists a \(y_0\) such that \(F(x_0, y_0) = 0\) in the range \(|y_0 - b| < \varepsilon\) (i.e., \(y_0 \in (b - \varepsilon, b + \varepsilon)\)). Moreover, there is exactly one such \(y_0\) as \(F(x_0,y)\) is strictly monotone increasing. Define the open interval \(I = (a - \varepsilon', a + \varepsilon')\). For each \(x_0 \in I\), we can determine a unique \(y_0\) as the (unique) solution of \(F(x_0,y_0) = 0\). In other words, we can define a function \(y = \varphi(x)\) in this way, and this function satisfies the conditions (1) and (2) of the Theorem.

Step 2. Showing that \(\varphi(x)\) is continuous on \(I\).

It suffices to show that, for all \(x_0\in I\), \(\varphi(x_0 + h) \to \varphi(x_0)\) as \(h \to 0\). We prove this by contradiction. 

Suppose \(\lim_{h\to 0}\varphi(x_0 + h) \neq \varphi(x_0)\). Then we can find a sequence \({h_n}\) that converges to 0 (\(\lim_{n\to\infty}h_n = 0\)) and a positive real number \(\varepsilon'' > 0\) such that
  • For all \(n \in \mathbb{N}\), \(|\varphi(x_0 + h_n) - \varphi(x_0)| \geq \varepsilon''\).
By construction in Step 1, for all \(n\in\mathbb{N}\), \(\varphi(x_0 + h_n) \in [b - \varepsilon, b + \varepsilon]\). Thus, by the Bolzano-Weierstrass Theorem:
"A sequence that is bounded in a closed interval contains a subsequence that converges to some value in the interval."
we can find a subsequence \(\{h_{n_k}\}\) of \(\{h_n\}\) such that \(\lim_{k\to\infty}\varphi(x_0 + h_{n_k}) = \alpha \in [b - \varepsilon, b + \varepsilon]\) where \(\alpha\) is a constant. Since \(|\varphi(x_0 + h_{n_k}) - \varphi(x_0)| \geq \varepsilon''\) for all \(k\), we have \(|\alpha - \varphi(x_0)|\geq \varepsilon''\) and, in particular, \(\alpha \neq \varphi(x_0)\).

Now, consider a point sequence on \(\mathbb{R}^2\): \(\{(x_0 + h_{n_k}, \varphi(x_0 + h_{n_k}))\}\). This point sequence converges to \((x_0, \alpha)\). Since \(F(x,y)\) is continuous, \(F(x_0 + h_{n_k}, \varphi(x_0 + h_{n_k}))\) converges to \(F(x_0, \alpha)\) as \(k \to \infty\). However, for all \(k\), we have \(F(x_0 + h_{n_k}, \varphi(x_0 + h_{n_k})) = 0\) and, in particular, \(F(x_0, \alpha) = 0\). We also have \(F(x_0, \varphi(x_0)) = 0\) by construction. On the other hand, if any \(y_0 \in [b - \varepsilon, b + \varepsilon]\) satisfies \(F(x_0, y_0) = 0\), it must be unique (for a given \(x_0\)). Therefore, it must be that \(\alpha = \varphi(x_0)\), which is a contradiction. Thus, \(\varphi(x)\) is continuous on \(I\).

Step 3. Showing that \(\varphi(x)\) is differentiable on \(I\).

It suffices to show that \(\lim_{h\to 0}\frac{\varphi(x + h) - \varphi(x)}{h}\) exists. By Taylor's Theorem, we can write
\[F(x + h, y + k) = F(x,y) + F_x(x+\theta h, y + \theta k)h + F_y(x+\theta h, y + \theta k)k\]
for some \(\theta \in (0, 1)\).  (See also: Taylor's theorem and asymptotic expansion)
If we set \(y =\varphi(x)\) and \(k = \varphi(x + h) - \varphi(x)\), then, since \(F(x, \varphi(x)) = F(x + h, \varphi(x + h)) = 0\), we have
\[F_x(x + \theta h, \varphi(x) + \theta k)h + F_y(x + \theta h, \varphi(x) + \theta k)k = 0,\]
that is,
\[\frac{\varphi(x + h) - \varphi(x)}{h} = \frac{k}{h} = -\frac{F_x(x + \theta h, \varphi(x) + \theta k)}{F_y(x + \theta h, \varphi(x) + \theta k)}.\]
Since \(\varphi(x)\) is continuous, we have \(k \to 0\) as \(h \to 0\). By the continuity of \(F_x(x,y)\) and \(F_y(x,y)\), we have
\[\lim_{h\to 0}\frac{\varphi(x + h) - \varphi(x)}{h} = -\frac{F_x(x,\varphi(x))}{F_y(x,\varphi(x))}.\tag{Eq:dphi}\]
Therefore, \(\varphi(x)\) is differentiable on \(I\) and its derivative is given by the right-hand side of (Eq:dphi).
■

Comments

Post a Comment

Popular posts from this blog

Open sets and closed sets in \(\mathbb{R}^n\)

-
Image
Open sets In \(\mathbb{R}\), we have the notion of an open interval such as \((a, b) = \{x \in \mathbb{R} | a < x < b\}\). We want to extend this idea to apply to \(\mathbb{R}^n\). We also introduce the notions of bounded sets and closed sets in \(\mathbb{R}^n\). Recall that the \(\varepsilon\)-neighbor of a point \(x\in\mathbb{R}^n\) is defined as \(N_{\varepsilon}(x) = \{y \in \mathbb{R}^n | d(x, y) < \varepsilon \}\) where \(d(x,y)\) is the distance between \(x\) and \(y\). Definition (Open set) A subset \(U\) of \(\mathbb{R}^n\) is said to be an  open set  if the following holds: \[\forall x \in U ~ \exists \delta > 0 ~ (N_{\delta}(x) \subset U).\tag{Eq:OpenSet}\] That is, for every point in an open set \(U\), we can always find an open ball centered at that point, that is included in \(U\). See the following figure. Perhaps, it is instructive to see what is  not  an open set. Negating (Eq:OpenSet), we have \[\exists x \in U ~ \forall \delta > 0 ~ (N_{\delta}(x) \not
Read more

Euclidean spaces

-
Image
We would like to study multivariate functions (i.e., functions of many variables), continuous multivariate functions in particular. To define continuity, we need a measure of "closeness" between points. One measure of closeness is the Euclidean distance. The set \(\mathbb{R}^n\) (with \(n \in \mathbb{N}\)) with the Euclidean distance function is called a Euclidean space. This is the space where our functions of interest live. The real line is a geometric representation of \(\mathbb{R}\), the set of all real numbers. That is, each \(a \in \mathbb{R}\) is represented as the point \(a\) on the real line. The coordinate plane , or the \(x\)-\(y\) plane , is a geometric representation of \(\mathbb{R}^2\), the set of all pairs of real numbers. Each pair of real numbers \((a, b)\) is visualized as the point \((a, b)\) in the plane. Remark . Recall that \(\mathbb{R}^2 = \mathbb{R}\times\mathbb{R} = \{(x, y) | x, y \in \mathbb{R}\}\) is the Cartesian product of \(\mathbb{R}\) with i
Read more

Newton's method

-
Image
Newton's method (the Newton-Raphson method) is a very powerful numerical method for solving nonlinear equations.  Suppose we'd like to solve a non-linear equation \(f(x) = 0\), where \(f(x)\) is a (twice) differentiable non-linear function. Newton's method generates a sequence of numbers \(c_1, c_2, c_3, \cdots\) that converges to a solution of the equation. That is, if \(\alpha\) is a solution (i.e., \(f(\alpha) = 0\)) then,  \[\lim_{n\to\infty}c_n = \alpha,\] and this sequence \(\{c_n\}\) is generated by a series of linear approximations of the function \(f(x)\). Theorem (Newton's method) Let \(f(x)\) be a function that is twice differentiable on an open interval \(I\) that contains the closed interval \([a, b]\) (i.e., \([a,b]\subset I\)) and satisfy the following conditions: \(f(a) < 0\) and \(f(b) > 0\); For all \(x\in [a, b]\), \(f'(x) > 0\) and \(f''(x) > 0\). Let us define the sequence \(\{c_n\}\) by \[ \begin{eqnarray} c_1 &
Read more