# Is a non-perfect improvement and optimisation?

In real word problems, the influence of multiple not perfectly known factors results in using heuristics instead of mathemacial solutions that calculates a perfect value from only precisly defined input data. Consequently, any method that does not supply the mathematical maximum or minimum is not an optimisation but an improvement.

Somehow my opinion on this topic differs from the use of the term optimisation in many papers. Are the people just not precise in their language or is my understanding of the term wrong?

Improvement doesn't sound as facy as optimisation, but is there maybe some facy word that allows people to still be precise?

• This is sort of a linguistic, maybe philosophical, question. Does "optimal" mean "best possible" or "best available"? A more practical answer might be that "optimization" in this context is understood to be subject to some fixed constraints, whether they've been properly defined or not, in which case this is true optimization in the sense you intend. – Patrick87 Oct 14 '13 at 15:23
• If one includes time/effort to solution (possibly over several different applications), limited knowledge (of inputs and of algorithms), reliability/accuracy of the solution (e.g., avoiding bad solutions given somewhat inaccurate inputs), flexibility of application (e.g., knowledge changing before solution is used or reuse to reduce later time to solution), and other real-world factors, then such might be considered "true" optimization; but usually the term is used as a synonym for improvement (but with emphasis on the action rather than just describing the effect--not merely being 'fancy'). – Paul A. Clayton Oct 14 '13 at 15:54