基于线性规划和组合理论的饲料多配方平行设计的探讨

Probing into a Parallel Design with Multiple-Formula for Diets for Livestock and Poultry on the Base of Linear Programming and Combinatorial Theory

多配方平行设计是另一类日粮配合方法,本研究旨在应用这种方法到畜禽日粮配合.以线性规划和组合理论为基础,采用全段、小段和点段设计方案,同时配合35种和14种不同生长和生产阶段的猪、鸡日粮,比较与分析日粮单位价格、原料用量和养分含量.通过这种设计方法产生的119种猪、鸡日粮配方的能量、钙、磷、粗蛋白、氩基酸、总氮和矿物质的养分含量,都达到或超过美国国家研究委员会(NRC)提出的猪、鸡营养需要的推荐量,也满足最低成本的要求.实例表明,多配方平行设计能同时配合多个畜禽日粮和获得一种日粮的多个配方,并能在一个大的范围内选到更好的配方,应用于动物饲料配方设计具有实际可行性.

A parallel design with multiple-formula is another kind of method of formulating diets. This study was conducted to apply the parallel design with multiple-formula to diet formulation for livestock and poultry. The method is based on linear programming and combinatorial theory, its three scenarios, including whole, subsection, and single point parallel designs,were used for formulating 35 diets for different growing and productive purposes of swine and 14 diets for chicken, and the amounts of feed Ingredients, nutrient levels and cost of the diets were compared and analyzed. Dietary energy, calcium, phosphorus, crude protein, amino acid, total nitrogen and mineral meet or exceed the nutrient requirement amounts of swine recommended by subcommittees on swine and poultry nutrition, national research council(NRC), and also the cost of the diets reach the minimum objective. The illustrations show that the parallel design with multiple-formula is able to simultaneously formulate multiple diets for livestock and poultry and generate each diet with multiple recipes, and to select a better least-cost formula in a large range,and has the actual feasibility for the animal feed formulation.