In the general "reproduction", a individual is choosen from the population, preferably a individual with an high fitness, than it is copied and modified by one operation. At the end it is added to the population. It can also be checked, if it is a "new" object (if there is no equal object in the population) and just added in this case, so there are no duplicate individuals in the population.
In the algorithm the reproduction is implemented by running an operator. This operation retrieves all the required data with the help of the operation interface of the algorithm. This required data may include one or more individuals or data on the recent course of the algorithm (e. g. the number of previous interactions respectively operations).