A general problem in model selection is to obtain the right parameters that make a model "t observed data. For a multilayer perceptron (MLP) trained with back-propagation (BP...
Pedro A. Castillo Valdivieso, Juan J. Merelo Guerv...
One problem of propagating the globally fittest individual via neighbourhood evolving in both island model and cellular model of existing parallel genetic algorithms (PGA) is that...
Many black box optimization algorithms have sufcient exibility to allow them to adapt to the varying circumstances they encounter. These capabilities are of two primary sorts: 1) ...
In this study, a new mechanism that adapts the mutation rate for each locus on the chromosomes, based on feedback obtained from the current population is proposed. Through tests us...
The paper presents the parameter-less implementation of an evolutionary-based search. It does not need any predefined control parameters values, which are usually used for geneti...