Extreme values of multivariate functions

-


A local maximum (or minimum) value of a function is the maximum (or minimum) value of the function in the neighbor of a point. More formally,

Definition (Local maximum, local minimum)

Let \(f(x,y)\) be a function on an open region \(U\subset \mathbb{R}^2\) and \(P=(a,b) \in U\).

  1. \(f(a,b)\) is said to be a local maximum value of the function \(f(x,y)\) if there exists \(\delta > 0\) such that, for all \((x,y)\in U\), if \((x,y)\in N_{\delta}(P)\cap U\) and \((x,y) \neq (a,b)\), then \(f(x,y) < f(a,b)\).
  2. \(f(a,b)\) is said to be a local minimum value of the function \(f(x,y)\) if there exists \(\delta > 0\) such that, for all \((x,y)\in U\), if \((x,y)\in N_{\delta}(P)\cap U\) and \((x,y) \neq (a,b)\), then \(f(x,y) > f(a,b)\).
Local maximum and minimum values are collectively called extreme values.

Theorem (Necessary condition for extreme values)

Let \(f(x,y)\) be a function on an open region \(U\), and \((a,b)\in U\). Suppose that \(f_x(a,b)\) and \(f_y(a,b)\) exist. If \(f(a,b)\) is an extreme value, then \(f_x(a,b) = f_y(a,b) = 0\).
Proof. If \(f(x,y)\) has an extreme value at \((a,b)\), then the univariate function \(f(x,b)\) of \(x\) has an extreme value at \(x = a\). By the corresponding theorem for univariate functions, \(f_x(a,b) = 0\). Similarly, \(f_y(a,b) = 0\). ■

The converse of this theorem is not necessarily true.

Example. Consider \(f(x,y) = x^2 - y^2\) (See Fig:NoExtreme below). We can see that \(f_x(0,0) = f_y(0,0) = 0\). However, the graph of \(f(x,0) = x^2\) is convex whereas that of \(f(0, y) = -y^2\) is concave. This means that, if we move from the origin along the \(x\)-axis, the value of \(f(x,y)\) increases, but if we move from the origin along the \(y\)-axis, the value decreases. Therefore, \(f(0,0)\)a  is neither a a local maximum nor a local minimum. □
Figrue Fig:NoExtreme. The graph of \(z=x^2 - y^2\). This function has neither a local maximum nor a local minimum at \((x,y) = (0,0)\). (See the above example.)

In the case of univariate functions, we can determine whether a function takes a local maximum or local minimum value at a given point by the sign of the second differential coefficient at that point. We have a corresponding theorem for the case of bivariate functions, but the situation is a bit more complicated. We remark the following theorem without proof.

Theorem (Criteria of extreme values)

Let \(f(x,y)\) be a function of class \(C^2\) on an open region \(U\). Suppose that \((a,b)\in U\) and \(f_x(a,b) = f_y(a,b) = 0\). Let us define the determinant \(D\) by
\[D = f_{xx}(a,b)f_{yy}(a,b) - [f_{xy}(a,b)]^2.\]
  1. If \(D > 0\),
    1. if \(f_{xx}(a,b) > 0\), then \(f(x,y)\) has a local minimum value at \((x,y) = (a,b)\);
    2. if \(f_{xx}(a,b) < 0\), then \(f(x,y)\) has a local maximum value at \((x,y) = (a,b)\).
  1. If \(D< 0\), \(f(x,y)\) does not have a local extremum at \((x,y) = (a,b)\).

Proof. Omitted (beyond the scope of this module). ■

Remark. Note that the "determinant" \(D\) in the above theorem is indeed the determinant of a matrix:\[D = \begin{vmatrix}f_{xx}(a,b) & f_{xy}(a,b)\\f_{yx}(a,b) & f_{yy}(a,b)\end{vmatrix}.\] This matrix is called the Hessian of the function. You may guess the Hessian of a general multivariate function \(f(x_1, x_2, \cdots, x_n)\). □

Remark. This theorem is only applicable to bivariate (two-variable) functions. For general multivariate functions, such as \(f(x_1, x_2, \cdots, x_n)\), the corresponding theorem is even more complicated (but it exists). That is, if all the eigenvalues of the Hessian are positive (or negative) at a given point, then the function takes a local minimum (or maximum) value at the point. (Note that the Hessian is always symmetric if the function is of class \(C^2\). Therefore, all the eigenvalues are guaranteed to be real.) □

Example. Let us find the extreme values of the function \(f(x,y) = x^3 + y^3 - 3(x+y)\) on \(\mathbb{R}^2\) (See the figure below).
The graph of \(z = x^3 + y^3 - 3(x+y)\).

Solving 
\[\begin{eqnarray*} f_x(x,y) &=& 3(x^2 - 1) = 0,\\ f_y(x,y) &=& 3(y^2 - 1) = 0, \end{eqnarray*}\]
we have
\[(x,y) = (1, 1), (-1, -1), (-1, 1), (1, -1).\]

From
\[\begin{eqnarray*} f_{xx}(x,y) &=& 6x,\\ f_{xy}(x,y) &=& 0,\\ f_{yy}(x,y) &=& 6y,\\ \end{eqnarray*}\]
we have 
\[D = f_{xx}(x,y)f_{yy}(x,y) - [f_{xy}(x,y)]^2 = 36xy.\]
  • When \((x,y) = (1, 1)\), we have \(D = 36 > 0\) and \(f_{xx}(1,1) = 6 > 0\) so that \(f(1,1) = -4\) is a local minimum.
  • When \((x,y) = (-1, -1)\), we have \(D = 36 > 0\) and \(f_{xx}(-1,-1) = -6 < 0\) so that \(f(-1,-1) = 4\) is a local maximum.
  • When \((x,y) = (-1, 1)\) or \((1, -1)\), we have \(D = -36 < 0\) so \(f(x,y)\) does not take extreme values.
□




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