Services on Demand
Journal
Article
Indicators
- Cited by SciELO
- Access statistics
Related links
- Cited by Google
- Similars in SciELO
- Similars in Google
Share
Revista científica
Print version ISSN 0124-2253On-line version ISSN 2344-8350
Abstract
LLANOS-MOSQUERA, José-Miguel; MURIEL-LOPEZ2, Gerardo-Luis; TRIANA-MADRID, Joshua-David and BUCHELI-GUERRERO, Víctor-Andrés. Evolutionary Algorithms Guided by Scale-Free Complex Networks. Rev. Cient. [online]. 2022, n.44, pp.228-241. Epub July 08, 2022. ISSN 0124-2253. https://doi.org/10.14483/23448350.18039.
Evolutionary computation algorithms allow solving optimization problems through defined iterations and stages. One of the most commonly employed techniques for this type of problem is differential evolution, which contains properties of small-world complex networks, whose study is important because of the results they generate for optimization problems. Considering the results obtained in previous works, which propose an evolutionary algorithm guided by complex small-world networks, a proposal is defined which contains complex scale-free networks, with the purpose of validating the averages generated by complex networks against the results obtained by the traditional evolutionary algorithm. An experiment was defined which allows evaluating the performance of the proposed model and that of the evolutionary algorithm by means of statistic indicators. Four optimization problems (Ackley, Beale, Camel, and Sphere) were also used to evaluate the hypothesis in the proposed model, its convergence, and the reduction of execution times compared to the base model. It was observed that the scale-free complex networks generated better averages than the traditional evolutionary algorithm and the small-world networks because they use a connection preferential mechanism between their nodes and guide the combination of individuals (solutions), thus improving the convergence rate and the performance of the evolutionary algorithm in general.
Keywords : Barabasi model; complex networks; population dynamics; scale-free network; traditional evolutionary computation..