Meta-Theory of Optimization
Optimization is the discipline (art/science) which is concerned with finding the maxima and minima of functions, subject to constraints.
In Meta-Terms, Optimization does not necessarily mean finding the optimum solution to a problem. Often this is not possible (constraints), and this is where the balance (heuristics) is used to make principally accurate judgments factoring in given circumstances. Then these are tested and refined to a degree that is itself part of the Optimization, which grows into a self-referential diminishing curve.
Knowing what questions to ask (your metrics), what tests to run (your data), and what to do with the information (analysis) is the Art of Optimization.
This is a definition in progress. What do you think? Let me know