The Simple Genetic Algorithm Performance: A Comparative Study on the Operators Combination
Delmar Broglio Carvalho, João Carlos N. Bittencourt, Thiago D'Martin Maia, Benchmarking Functions · 2011
This paper presents a comparative and experimental study about the performance of the Simple Genetic Algorithm (SGA) using five classic benchmarking functions. The performance analysis is accomplished on the combination of the operators of reproduction and crossover with the control parameters having been fixed. The overall behavior of the SGA is evaluated by the fitness of the best individual analyzed during the evolution and at the ending of the same one. The results that are presented show that the SGA can be effective and competitive to optimization on a test suite of benchmark functions. Keywords-Genetic Algorithm; Parameterization of GA; Generational Replacement Model; Single-Point Crossover; Uniform Crossover.