Average Trends over Millennia of Evolution Supervised by Genetic Algorithms. Analysis of Phenotypes

Abstract: A genetic algorithm (GA) had been developed and implemented in order to identify the optimal solution in term of determination coefficient and estimation power of a multiple linear regression approach on structure-activity relationships. The Molecular Descriptors Family was used for structure characterization of a sample of 206 polychlorinated biphenyls with measured octanol-water partition coefficients as case study. The research aimed to analyze the degree of association between the number of viable genotypes in cultivar in generations when the evolution occurred and the selection and survival strategies. The GA was run by 46 times and the Anderson-Darling test was used to compared the distribution laws of populations occurred by using different pairs of survival and selection strategies. The records of genotypes number were analyzed and the conclusions were highlighted