The Optimization Analysis Of Mixture Ratio And Product Mixture Design Based On Genetic Algorithms

There are many drug proportion optimization problems in the field of pharmaceutical research. Mixture experiment design is a design that optimization the function between the mixture index and the mixture components percentages, to determine the optimal proportion. The content of each mixture composition must be expressed as percentage, and the sum must be1.With the mixture constraints, how to design mixture experiment, how to determine the optimal proportion, to reduce costs and improve efficiency is an important issue in the research of the optimization of mixture experiments.The traditional optimization method is the contour method, but its subjectivity that limits its application. Genetic algorithm is a global random search optimization algorithm and can effectively avoid local optimization. Genetic algorithms demonstrates a greater advantage than traditional methods in mixture design with constraints.The topic of the research is to introduce mixture ratio, product mixture experiment design and genetic algorithm, and provids a feasible method with mixture design of optimization problem.The first part:The summarization of mixture experiment design. It introduces the concept, principle of mixture ratio and product mixture experiment design. Mixture ratio design:When it is p factors in the mixture problem, which the sum of the p factors isl, one of the factors can not vary independently. If the p factors are transformed into (p-1) ratio factors, then all the ratio factors can vary independently. Each ratio factors are no longer restricted by mixing constraints, in turn, can adopt the orthogonal design. Product mixture design:When it containes another type factors which different from the mixture factors in the mixture problems, it usually combines simplex centroid or simplex lattice design with the factors of two level factors design constructed experimental program. The second part:The genetic algorithm with the restriction of mixture conditions.Introduces the principle of genetic algorithm and vector evaluated genetic algorithm.The third part:The optimize effect of the mixture experiments design with genetic algorithm. Adopt the method of second order partial derivatives, contour method, genetic algorithm optimization. The solutions of second order partial derivative are beyond the constraints of test function. The contours method can not give precise solutions. Genetic algorithm can give precise selection of the appropriate combination of the independent variables, and the solutions are stable.The forth part:the optimized exploratory research with mixture design based on genetic algorithm.The mixture ratio design:(1)The optimized research of fly ash concrete mixture ratio design based on vector evaluated genetic algorithm for multi-objective optimization, the optimal proportion of new ingredients:slag41.95%, coal ash27.97%,water15.38%, sodium silicatel2.58%and sodium hydroxide2.09%,as the three days and twenty eight days compressive strength of mortar are32.16Pa and63.93Pa, and the solution are superior to contour method.(2)The optimized research of soybean oil compound enzyme legal system mixture ratio design based on genetic algorithm for optimization, the optimal proportion of the components:cellulase11.37%, pectinase62.34%and protease26.29%.The fat obtained under this condition is77.24%and the solutions are superior to original text.The product mixture design:The optimized research of coating on titanium substrate solution product mixture design based on genetic algorithm for optimization, the optimal solutions are:rubidium oxide52.4%, tin oxide47.6%, the process variables calcination temperature is500℃, the electrode life reached up to32.6h and the optimal value is superior to contour method.From the above, genetic algorithms suits for the optimization analysis of mixture ratio design and product mixture design with the constraint condition, providing the reasonable plan.