Challenges of Bioinspired Artificial Intelligence with Genetic Algorithms

Desafíos de la Inteligencia Artificial Bioinspirada con Algoritmos Genéticos

Abstract

At first, basic definitions are presented to understand the scope of Artificial Intelligence, highlighting their positions regarding the possibility of emulating human reasoning on a computer. Specifically, efforts are identified to define formal systems that can model mathematics and therefore human reasoning under the assumption that human intelligence functions under a mechanistic paradigm. However, there are insurmountable limitations in all formal systems so we must settle for strategies that converge to sub-optimal solutions. A current of Artificial Intelligence establishes this search through techniques inspired by biology, which go beyond the intrinsic limits of deductive mathematical reasoning, since they have proven their effectiveness in nature itself. Consequently, the basic principles that govern evolution and genetics, and the dynamics that underlie genetic algorithms in general, are presented. Finally, some results that we have published will be shown, using in particular mono-objective and multi-objective genetic algorithms, to solve problems to which, in principle, formal operations research techniques or formal systems in general cannot be applied.