A Linear Time Algorithm for Determining Population Diversity in Evolutionary Computation

M. Wineberg and F. Oppacher (Canada)


Genetic Algorithms, Evolutionary Computation, Diversity


In this paper we analyze the most common diversity measures used in Evolutionary Computation, the sum of all-possible-paired distances, which can be used for binary, symbolic and real valued genomes. We present anO(n) implementation for this family of diversity measures, despite its basis in an O(n 2 ) number of comparisons. Finally, we present other common diversity measures and compare and contrast them with the all possible-pairs diversity.

Important Links:

Go Back