Use of “Controlled Random Search Technique for Global Optimization” in Animal Diet Problem

Abstract—Linear programming has been used widely in practice with noticeable success for formulating optimum livestock diets. Diet formulated by linear programming is based on the assumption of linearity between animal yield and nutrient ingredients included in the diet. To overcome the drawback of linear approximation of objective function for diet formulation, there is a need to look for techniques which can handle linearity as well as non-linearity to a large extent with acceptable solutions. The present work represents a departure from the traditional LP approach by solving the diet formulation problem using the Controlled Random Search Technique (RST2) for global optimization. This technique is probabilistic in nature and does not take into account the mathematical nature of the functions but at the same time gives promising results. This study is aimed at comparing the optimal solution of linear model of animal diet formulation by Excel solver and RST2.