Viveros Jiménez F., Mezura-Montes E., Gelbukh A. (2009) Elitistic Evolution: An Efficient Heuristic for Global Optimization. In: Kolehmainen M., Toivanen P., Beliczynski B. (eds) Adaptive and Natural Computing Algorithms. ICANNGA 2009. Lecture Notes in Computer Science, vol 5495. Springer, Berlin, Heidelberg
A new evolutionary algorithm, Elitistic Evolution (termed EEv), is proposed in this paper. EEv is an evolutionary method for numerical optimization with adaptive behavior. EEv uses small populations (smaller than 10 individuals). It have an adaptive parameter to adjust the balance between global exploration and local exploitation. Elitism have great influence in EEv’ proccess and that influence is also controlled by the adaptive parameter. EEv’ crossover operator allows a recently generated offspring individual to be parent of other offspring individuals of its generation. It requires the configuration of two user parameters (many state-of-the-art approaches uses at least three). EEv is tested solving a set of 16 benchmark functions and then compared with Differential Evolution and also with some well-known Memetic Algorithms to show its efficiency. Finally, EEv is tested solving a set of 10 benchmark functions with very high dimensionality (50, 100 and 200 dimensions) to show its robustness.