Adaptive Genetic Algorithm Based on Crossover and Mutation Method for Optimization of Poultry Feed Composition

Abstract

The problem that often comes in poultry industry is the process of composition selection of poultry feed that is not efficient. Feed ingredients with high nutrients have a relatively expensive price while low-cost feeds contain inadequate nutritional needs of livestock. Breeders should reduce the cost of spending to generate higher income. These problems can be solved using genetic algorithms to obtain the composition of feed ingredients in accordance with the needs of livestock nutrition and have a minimum price. The simple genetic algorithm applied takes a long time to reach the optimal solution. Therefore, genetic operator should be improved to increase the ability of genetic algorithm in finding the optimal solution. This study uses a combination of different crossover and mutation methods. We propose extended intermediate and one cut point method in the crossover process. We use two method of mutation which are random and inverse mutation. The experiment result shows that the optimum population size is 250, the best combination of cr and mr of this experiment is 0.7 and 0.3 and the number of generations is 1000. According to the result, the experiment proves that adaptive genetic algorithm is efficient in solving this kind of problems.