The effectiveness of mutations: Simulation of natural mechanisms

Current versions of Genetic Algorithms make use of point mutation only, in spite of the existence in nature of many different forms of mutations. In this paper we present an Evolutionary Algorithm based on the simulation of natural mutation mechanisms acting at gene level and chromosome level. Experimental trials are effected on several test functions. The effectiveness of the approach proposed is demonstrated through a comparison with the results obtained by classical crossover-based Genetic Algorithms.