Reduction of fitness evaluations using clustering technique and neural network ensembles

摘要

The invention proposes an evolutionary optimization method. In a first step, an initial population of individuals is set up and an original fitness function is applied. Then the offspring individuals having a high evaluated quality value as parents are selected. In a third step, the parents are reproduced to create a plurality of offspring individuals. The quality of the offspring individuals is evaluated by means of a fitness function, wherein selectively the original or an approximate fitness function is used. Finally, the method goes back to the selection step until a termination condition is met.;The step of evaluating the quality of the offspring individuals consists in grouping all λ offspring individuals in clusters, selecting for each cluster one or a plurality of offspring individuals, resulting in altogether ξ selected offspring individuals, evaluating the ξ selected offspring individuals by means of the original fitness function, and evaluating the remaining λ-ξ offspring individuals by means of the approximate fitness function.