Optimization (mathematics)

In mathematics, the term optimization refers to the study of problems that have the form

Given: a function f : A -> R from some set A to the real numbers

Sought: an element x₀ in A such that f(x₀) ≥ f(x) for all x in A ("maximization") or such that f(x₀) ≤ f(x) for all x in A ("minimization").

Such a formulation is called a mathematical program. A great many real-world and theoretical problems may be modeled in this general framework.

Typically, A is some subset of Euclidean space Rⁿ, often specified by a set of constraints, equalities or inequalities that the members of A have to satisfy. The elements of A are called the feasible solutions and the function f is called the objective function. A feasible solution that maximizes (or minimizes, if that is the goal) the objective function is called an optimal solution.

In general there will be several local maxima and minima, where a local minimum x^* is defined as as a point such that for some δ > 0 and all x such that ||x - x^* || ≤ δ the formula f(x) ≥ f(x^*) holds; that is to say on some ball around x^* all of the function values are greater than the value at that point. Local maxima are defined similarly. In general, it is easy to find local minima, however additional facts about the problem (e.g. the function being convex) are required to ensure that the solution found is a global minimum.

Table of contents

1 Notation
2 Techniques
3 Uses
4 History

Notation

Optimization problems are often expressed with special notation. Here are some examples:

min_{x in R} x²+1

This asks for the minimum value for the expression x²+1, where x ranges over the real numbers R. The answer in this case is 1.

max_{x in R} 2x

This asks for the maximum value for the expression 2x, where x ranges over the reals. In this case, there is no such maximum as the expression is unbounded, so the answer is "infinity" or "undefined".

arg min_{x in ]-∞,-1]} x²+1

This asks for the value(s) of x in the interval ]-∞,-1] which minimizes the expression x²+1. (The actual minimum value of that expression does not matter.) In this case, the answer is x = -1.

arg max_{x in ]-∞,5], y in R} x · cos(y)

This asks for the (x,y) pair(s) that maximize the value of the expression x·cos(y), with the added constraint that x cannot exceed 5. (Again, the actual maximum value of the expression does not matter.) In this case, the solutions are the pairs of the form (5,2&pik) and (-5,(2k+1)π), where k ranges over all integers.

Techniques

Techniques for solving mathematical programs depend on the nature of the objective function and constraint set. The following major subfields exist:

linear programming studies the case in which the objective function f is linear and the set A is specified using only linear equalities and inequalities
integer programming studies linear programs in which some or all variables are constrained to take on integer values
quadratic programming allows the objective function to have quadratic terms, while the set A must be specified with linear equalities and inequalities
nonlinear programming studies the general case in which the objective or constraints or both contain nonlinear parts
stochastic programming studies the case in which some of the constraints depend on random variables
dynamic programming studies the case which has optimal substructure and overlapping subproblems.

For derivable functions, unconstrained problems can be solved by finding the places where the gradient of the function is 0 (i.e. the stationary points) and using the Hessian to classify the type of point. If the hessian is positive definite, the point is a local minimum, if negative definite, a local maximum, and if indefinite it is some kind of saddle point.

Should a function be convex over a region of interest (as defined by constraints) then any local minimum will also be a global minimum. Robust, fast numerical techniques exist for optimizing doubly differentiable convex functions. Outside of these functions, less ideal techniques must be used.

Constrained problems can often be transformed into unconstrained problems with the help of the Lagrange multiplier.

Several techniques exist for find a good local minimum in nonlinear optimization problems with many poor local minima:

Uses

Additionally, problems in rigid body dynamics (in particular articulated rigid body dynamics) often require mathematical programming techniques, since you can view rigid body dynamics as attempting to solve an ordinary differential equation on a constraint manifold; the constraints are various nonlinear geometric constraints such as "these two points must always coincide", "this surface must not penetrate any other", or "this point must always lie somewhere on this curve". Also, the problem of computing contact forces can be done by solving a linear complementarity problem, which can also be viewed as a QP (quadratic programming problem).

(Are there other fields with strong connections? Economics)

History

Historically, the first term to be introduced was linear programming, which was invented by George Dantzig in the 1940s. The term programming in this context does not refer to computer programming (although computers are nowadays used extensively to solve mathematical programs). Instead, the term comes from the use of program by the United States military to refer to proposed training and logistics schedules, which were the problems that Dantzig was studying at the time. (Additionally, later on, the use of the term "programming" was apparently important for receiving government funding, as is was associated with high-technology research areas that were considered important.)

In fact, some mathematical programming work had been done previously...(anyone? - Gauss did some stuff here)

See also:

External links:

NEOS Guide