Research On Application Of Genetic Algorithm In Swine Feed Formula System

Feed is the material basis of breeding industry and high quality feed can promotethe benign development of aquaculture. There are usually many complex calculations infeed formula designing. Manual computational method cannot satisfy the actual needsand the compute computational method can only resolve the problem which has fewerlimits in small scale right now. In the studied area, genetic algorithm has been applied infeed formula system. Because genetic algorithm can handle more factors and high degreenonlinear problem, it can better conform to the nutritional content standards.In this thesis, we will do some research on the application in feed formula system.There are two standards to evaluate the feed formula, one is the economic costs and theother is the ammonia standard. If ammonia is excess, it can make animals cause diseasesand it is harmful to the natural environment. Therefore, in addition to establish areasonable ventilation system such as the improvement of the hardware environment, weshould also consider to optimize the feed formula, improve nutrition levels and reduceammonia emissions.In this paper, we will use the linear programming method, the standard geneticalgorithm and the NSGA-II algorithm to solve the problems of the pig feed formula andwe also will do some analysis and comparisons among the three results. The linearprogramming method will be realized by using Lingo software; we will adopt Matlabgenetic algorithm toolbox to accomplish the implementation of the standard geneticalgorithm. At last, we will do some programming to finish the NSGA-II algorithm.NSGA-II algorithm is a rapid non dominated sorting multi-objective genetic algorithmmainly used in the multi-objective optimization problem. The research emphasis of thispaper is the application of NSGA-II algorithm in pig feed formula system. At first, weshould transfer feed formulation to multi-objective optimization problem; it means the economic cost and ammonia emissions should be converted into the correspondingobjective function. And then, we will make some improvements to the crossover strategy,mutation strategy and Pareto sorting strategy of the NSGA-II algorithm. At last, we dosome programming using c++language. According to the contrastive analysis of theexperimental results, we can draw a conclusion that the NSGA-II algorithm is a goodselection to solve the problem of feed formulation system designing. It can red uce theeconomic cost of feed as well as ammonia emissions.