In essence, an optimization has three parts:
- optimization parameters,
- a cost function, and
- (optionally) constraints.
The optimization problem is to find the parameters that minimize the cost function, subject to any constraints given. In math, this is written as
min J(x)
where x are the optimization parameters (a d-dimensional vector), is the cost function, and D represents the constraints. If the problem is unconstrained, then D=R . Examples of constraints include those that enforce positivity, or x must satisfy some other set of equations or inequalities.