# Calculus Optimization Methods

A key application of calculus is in optimization: finding maximum and minimum values of a function, and which points realize these extrema.

### Context

Formally, the field of mathematical optimization is called mathematical programming, and calculus methods of optimization are basic forms of nonlinear programming. We will primarily discuss finite-dimensional optimization, illustrating with functions in 1 or 2 variables, and algebraically discussing n variables. We will also indicate some extensions to infinite-dimensional optimization, such as calculus of variations, which is a primary application of these methods in physics.

### Techniques

Basic techniques include the first and second derivative test, and their higher-dimensional generalizations.

A more advanced technique is Lagrange multipliers, and generalizations as Karush–Kuhn–Tucker conditions and Lagrange multipliers on Banach spaces.

### Applications

Optimization, particularly via Lagrange multipliers, is particularly used in the following fields:

Further, several areas of mathematics can be understood as generalizations of these methods, notably Morse theory and calculus of variations.

### Terminology

• Input points, output values
• Maxima, minima, extrema, optima
• Stationary point, critical point; stationary value, critical value
• Objective function
• Constraints – equality and inequality
• Especially sublevel sets
• Feasible region, whose points are candidate solutions

## Statement

This tutorial presents an introduction to optimization problems that involve finding a maximum or a minimum value of an objective function ${\displaystyle f(x_{1},x_{2},\ldots ,x_{n})}$  subject to a constraint of the form ${\displaystyle g(x_{1},x_{2},\ldots ,x_{n})=k}$ .

## Maximum and minimum

Finding optimum values of the function ${\displaystyle f(x_{1},x_{2},\ldots ,x_{n})}$  without a constraint is a well known problem dealt with in calculus courses. One would normally use the gradient to find stationary points. Then check all stationary and boundary points to find optimum values.

### Example

• ${\displaystyle f(x,y)=2x^{2}+y^{2}}$
• ${\displaystyle f_{x}(x,y)=4x=0}$
• ${\displaystyle f_{y}(x,y)=2y=0}$

${\displaystyle f(x,y)}$  has one stationary point at (0,0).

## The Hessian

A common method of determining whether or not a function has an extreme value at a stationary point is to evaluate the hessian of the function at that point. where the hessian is defined as

${\displaystyle H(f)={\begin{bmatrix}{\frac {{\partial }^{2}f}{\partial x_{1}^{2}}}&{\frac {{\partial }^{2}f}{\partial x_{1}\partial x_{2}}}&\dots &{\frac {{\partial }^{2}f}{\partial x_{1}\partial x_{n}}}\\{\frac {{\partial }^{2}f}{\partial x_{2}\partial x_{1}}}&{\frac {{\partial }^{2}f}{\partial x_{2}^{2}}}&\dots &{\frac {{\partial }^{2}f}{\partial x_{2}\partial x_{n}}}\\\vdots &\vdots &\ddots &\vdots \\{\frac {{\partial }^{2}f}{\partial x_{n}\partial x_{1}}}&{\frac {{\partial }^{2}f}{\partial x_{n}\partial x_{2}}}&\dots &{\frac {{\partial }^{2}f}{\partial x_{n}^{2}}}\\\end{bmatrix}}.}$

## Second derivative test

The Second derivative test determines the optimality of stationary point ${\displaystyle x}$  according to the following rules [2]:

• If ${\displaystyle H(f)>0}$  at point x then ${\displaystyle f}$  has a local minimum at x
• If ${\displaystyle H(f)<0}$  at point x then ${\displaystyle f}$  has a local maximum at x
• If ${\displaystyle H(f)}$  has negative and positive eigenvalues then x is a saddle point
• Otherwise the test is inconclusive

In the above example.

${\displaystyle H(f)={\begin{bmatrix}4&0\\0&2\end{bmatrix}}.}$

Therefore ${\displaystyle f(x,y)}$  has a minimum at (0,0).

## References

[1] T.K. Moon and W.C. Stirling. Mathematical Methods and Algorithms for Signal Processing. Prentice Hall. 2000.
[2]http://www.ece.tamu.edu/~chmbrlnd/Courses/ECEN601/ECEN601-Chap3.pdf