Environmental/survivors’ selection

In this phase, it is decided who will survive for the next generation/iteration. Obviously, the survival of good solutions will lead the algorithm to converge while it may cause the algorithm to converge prematurely. Hence, some poor solutions should be given chance to survive, which will allow GA to maintain diversity and enhance exploration. Many selection methods have been proposed in the literature. A few of them are talked about below:

Random selection

In this type of selection scheme, each solution regardless of its fitness will be given an equal chance to survive. Obviously, there will be a chance to lose good solutions.

Proportionate selection

This is exactly the same method that is discussed in the matting section.

Merging parents and new offspring

In such a case, parents and new offspring are merged to make the next generation. Usually, the number of offspring is decided by subtracting the matting pool size from the total population size.