Steps & Operations #2
Description
Subsequently, from #46 (closed) another concept that only exists in the analytics is "Steps" and "Operations".
We define a Step as something an algorithm implicitly does to get closer to a solution. Ex. a GA produces the next generation, or a LocalSearch attempts the next modification.
We define an Operation as something more generalizable than a step. This concept is not necessarily restricted to only one algorithm. It is also something that has strategic value, and there can be different options for a single algorithm. Ex. A Mutator would be an operation needed for virtually all (currently implemented) algorithms. Another example would be the different kinds of selection in a GA (Tournament, Fitness-Proportionate, ...).
While a step has little relevance besides making algorithms comparable by length using the analytics, operations have the added value of enabling us to answer questions such as "with mutation strategy has a higher impact on finding a good solution to problem x?".
Go and implement this. (fyi operations are already implemented in GCE).
Addendum: Another concept is "Population" where solutions are actually dependent on each other, as they are viewed as a whole. We might want to integrate this in ml as well.