An Evolutionary Algorithm Partially Inspired in the Vertebrates' Immune System for Performing Numerical Optimization

R.P. Hernández García (Mexico)

Keywords

Genetic algorithms, numerical optimization, constraint handling.

Abstract

The Immune System of Jawed Vertebrates (ISJV) is the most evolved of the immune systems [1], it has inspired the investigators in the development of solutions for problems in the field of computer sciences [2]. It is able to evolve inside the organism it defends in order to identify and mark progressively better the pathogens that may threaten the organism's health. A group of mechanism used by the ISJV to better its recognition of pathogens is called Affinity Maturation (AM). One of these known mechanisms is the Somatic Hypermutation (SH) of the B lymphocytes (specialized cells of ISJV). Some researchers believe that another mechanism: the Receptor Editing (RE) of the B lymphocytes also acts during AM [3]. In this paper, an algorithm inspired in AM with RE in combination with the crossover operator is proposed for making numerical optimization. The algorithm is tested with some constrained problems shown in [4] and one unconstrained minimax problem. The proposed algorithm worked well.

Important Links:



Go Back