In evolutionary optimization, selection is like a survival test for candidate solutions, deciding which ones get to 'reproduce.' Selection operators are the rules determining who passes this test. They might choose the fittest solutions (those solving the problem best) or sometimes include random or less fit ones for diversity. 

elitism: select N most fit individuals and copy them to next generation

roulette-wheel: probability of survival into next generation is proportional to indvididual's fitness

tournament...

In evolutionary optimization, a "fitness function" is like a scoring system that rates how good each candidate solution (or 'individual') is at solving the problem. Think of it as a judge in a competition, evaluating participants based on how well they meet certain criteria. The higher the score, the better the solution. This function is crucial because it guides the selection process, helping to decide which solutions should be kept, discarded, or combined to create the next generation of solutions.