Optimization

Optimization means to find the set of parameters that minimize the cost of a solution to a problem.

A Global Optimization is the optimal solution out of all possible solutions. [PSOAI]

A Local Optimization is a solution that may not be the global optimum but is good enough. [PSOAI]

• Exhaustive: Try all the combinations.
• Stochastic: Try a random subset and pick the best.
• Learning: Pick random solutions that improve as you go.

The first phase is to create a mathematical model of the system.

• Translate parameters for the solution to numbers.
• Represent solutions as vectors of parameters.
• Create a cost function that maps the global minimizing parameters to the optimal solution. [PSOAI]

A set of parameters that solves the problem with the least cost while still meeting constraints. [IIS]

The set of valid parameters (those that solve the problem no matter the cost). [IIS]

The subset of valid parameters that meet the constraints of the problem. [IIS]

A solution that meets all the constraints. [PSOAI]

A function that maps solutions to their cost (maps the best solution to the best outcome and the worst solution to the worst outcome).