Optimization of Effective Variables in Broiler Breeding by Integrated Mahalanobis - Taguchi System and Simulated Annealing Algorithm

There are complexities in today's systems because of the large number of influential factors and the complex relationships that exist between these factors. Therefore, appropriate tools have been developed to analyze complex systems, including the Mahalanobis-Taguchi system, which can make systems simpler and more understandable by extracting real relationships between variables. Thereby in this study, developed Mahalanobis-Taguchi systems was applied to reduce the number of effective variables for broiler breeding, which is one of the important protein sources in Iran. In the developed system, metaheuristic algorithms such as Population-Based Simulated Annealing were used. The results showed that the Population-Based Simulated Annealing algorithm has the ability to optimize this problem and it can reduce the number of variables according to the weights of the target function that they selected by the decision maker. Moreover, in this study, different weights were considered for the objective function. The results and the effect of different weights and the number of variables output from the algorithm were discussed. Based on the different weights for the objective function, the number of variables decreased from 35 to 10 in the first case, 11 in the second case, 10 in the third case and 21 variables in the fourth case