Uma solução de recomendações de acasalamentos baseada em algoritmos genéticos

Abstract

By 2050, food production is expected to increase by between 40% and 60% in order to maintain the food security of nearly 10 billion people. The primary sector is undergoing critical changes focused on large-scale production in response to current and upcoming demand. Among the man-made foods, meat stands out as the main source of animal protein. For breeding nuclei, multipliers and marketers, one of the stages of the beef production process is the choice of mating in search of offspring that respond to the producers' wishes. One of the tools for comparing superior or inferior characteristics among the animals at the moment of choice for mating is the evaluation of their expected progeny differences (DEPs). The main objective of this work is to improve the choice of mating based on the Expected Progeny Differences. The search for combinations of mates considering multiple genetic characteristics in the search for the maximization of results becomes an unfeasible problem to be treated in polynomial time, being necessary a heuristic strategy for its resolution. The present work presents a solution based on genetic algorithms (AG) that analyzing the main set of variables involved in the process will recommend mating to the producer. The analyzes of the experiments indicate a satisfactory behavior of GA, since the results obtained show the inclination to choose the animals with the highest DEP value based on the weights determined by the producer.