Optimization Applications Of Mixture Data Of The Sustained-Controlled Release Formulation Based On Hyperspherical Transformation And PCA-QIF

Objective:The regression model of mixture data of the sustained-release and controlled-release formulation is recommended with applying the methods of dimension-reduction approach through a hyperspherical transformation and quadratic inference functions based on principal component.After establishing the unconstrained model of mixture data,the genetic algorithm is proposed for solving the problem of multi-objective optimization of mixture formulation.The optimization results are compared with studies from the literature.The research offers a reasonable method for modeling and optimizing of mixture data in the sustained-release and controlled-release formulation.Methods:An exploratory study using the two types of mixture data which embedded with zero components and which without zero components was conducted.The experiment was designed using the cumulative release at different point time as independent variables,the properties of drug as dependent variables.We use the hyperspherical transformation resolving the unit-sum in mixture data and use principal component analysis resolving the multiple correlation among variables.Employs three different work correlation matrixes to establish equation with the quadratic inference functions and obtain a better model by AIC and BIC.NSGA-? is used for multi-objective optimization of mixture formulation.The optimization results are compared with studies from the literature.Results:In a case about optimization the formulations of the nimodipine matrix tablet which isn't embedded with zero components,after the properties of drug hyperspherical transformed,the principal component analysis is used and extract 7 principal components which can explain 99.98% of the variance.the QIF model adopting the exchangeable work correlation matrix type is statistically significant and the effect is good.Using a NSGA-? to obtain pareto optimal solutions : All solutions are within the prescribed scope and the 12 h cumulative release of multiple optimal solutions achieve 99%.When the properties of HPMC,Lactose and Sodium are 0.2649,0.6464 and 0.0887 respectively,the predicted cumulative release of 12 hcan reach to 99.60%.The predicted cumulative release of 3h,6h and 9h are 21.21%,50.66% and 77.60% respectively.The original literature obtain 5 soulutions by establishing Scheffé polynomial model and optimizing by response surface diagram method.The 12 h cumulative release of 5 solutions don't achieve 99%.When the properties of HPMC,Lactose and Sodium are 0.3458,0.4715 and 0.1627 respectively,the predicted cumulative release of 12 hcan reach to 99.60%whoseis 0.95 lower than this study.In a case about optimization the formulations of the metronidazole sustained-release films which embedded with zero components,after the properties of drug hyperspherical transformed,the principal component analysis is used and extract 10 principal components which can explain 99.97% of the variance.the QIF model adopting the exchangeable work correlation matrix type is statistically significant and the effect is good.Using a NSGA-? to obtain pareto optimal solutions : The 1~4days cumulative release of multiple optimal solutions are less than 2% with the release goal gap.In 30 optimal solutions,the 5 day cumulative release achieve 90%.When the properties of metronidazole,PCL,HPMCand GMSare 0.044,0.817,0.115 and 0.023 respectively,The predicted cumulative release of 1~5 days are 40.32%,55.31%,70.9%,85.08% and 92.13% respectively.The original literature obtain 7 soulutions by establishing Scheffé polynomial model.When the predicted cumulative release of 1 day and 4 day reach to the release standard,the gap between the predicted cumulative release of 2 day,3day and 5 day and the release standard is big.The average of 7 solution for predicted cumulative release are 47.41%,68.71%,81.48%,88.37% and 92.56% respectively and the gap with the best release is big.Conclusions:Hyperspherical transformation and quadratic inference functions are applied to structure models in analysis of mixture data of the sustained-release and controlled-release formulation.The NSGA-? multi-objective genetic algorithm is proposed for solving the problem of optimization.The optimal formulation can be obtained though inverse transform of hyperspherical transformation.The entire solution for the analysis of mixture data is reasonable and practicable.The methods can effectively solve the multi-objective optimization problem in mixture data of the sustained-release and controlled-release formulation,and it also can be extended to similar optimization applications of other sectors.