Hands-On Genetic Algorithms with Python

Rank-based selection

The rank-based selection method is similar to the roulette wheel selection, but instead of directly using the fitness values to calculate the probabilities for selecting each individual, the fitness is used just to sort the individuals. Once sorted, each individual is given a rank representing its position, and the roulette probabilities are calculated based on these ranks.

For example, let's take the same population of six individuals we previously used with the same fitness values. To that, we add the rank of each individual. As the population size in our example is six, the highest-ranking individual gets the rank value of 6, the next one gets the rank value of 5, and so on. The relative portion of the roulette wheel ...