02086nam a2200253 a 450000100080000000500110000800800410001910000290006024500840008926000760017330000140024949000350026350000750029852012780037365000280165165000130167965000110169265000180170365000160172170000220173770000310175970000230179070000190181321317172021-05-06       2019    bl uuuu     u00u1 u    #d1 aFONTOURA, D. de C. N. da  aOptimizing mate selectionba genetic algorithm approach.h[electronic resource]  aIn: ICAR CONFERENCE, 43., 2019, Prague. Proceedings... Rome: ICARc2019  ap. 54-62.  a(ICAR. Technical series n. 24)  aEditors:  J. Kucera, P. Bucek, D. Lipovsky, X. Bourrigan and M. Burke.  aBackground: Genetic Improvement Programs (GIP) aim to enhance productionefficiency of beef cattle. The main way to guide this enhancement is by choosing thebest mates among sires and cows, in order to maximize the offspring GeneticQualification Index (QGI), which is measured by an index defined by the GIP andcomputed for each animal of the herd. This paper describes a genetic algorithm, whichcan recommend an optimal set of matings among sires and cows, in order to maximizethe QGI of the herd. Breeders can define constraints regarding level of problems,which must be avoided, and they also can alter the traits relative importance consideredin QGI, according their particular interests. This algorithm was applied to a herd of aBrazilian breeder, which participates of a GIP, and it found optimal matings in order toincrease QGI value. We have simulated different scenarios considering variations onfitness functions, which combine QGI and level of problems, in order to find the optimalmatings. Proposed approach was successfully used to recommend optimal matingdecisions by Brazilian Hereford and Braford cattle breeders Association leading to animprovement of offspring QGI.Keywords: Genetic Improvement, Beef Cattle, Artificial Intelligence, EvolutionaryComputing.  aAcasalamento Controlado  aAnálise  aBovino  aGado de Corte  aPerformance1 aCAMARGO, S. da S.1 aTORRES JUNIOR, R. A. de A.1 aCARVALHO, H. G. de1 aCARDOSO, F. F.