| Objective:By changing the optimization strategy, transform the interval optimization to the point optimization, and using non-dominated sorting genetic algorithm-II(NSGA-II) optimize the lag time and the specified point cumulative release which have been transformed,and evaluating the effect of the method. Providing a more scientific and feasible statistical method to optimize the formulation of pulsatile drug delivery system. Methods:Putting each component of the pulsatile drug delivery system as the independent variable and putting the lag time and the specified point cumulative release as the dependent variable. Using the quadratic polynomial model to establish each objective function, and using the adjusted coefficient of determination(Ra2) as the evaluation indicator to judge the fitting effect. According to the requirements of the literature, the objective functions need to be transformed. non-dominated sorting genetic algorithm(NSGA-II) is used to optimize the objective functions that have been transformed. Select the optimal solutions to compare with the results of the literature. Results:In the optimization study of ranitidine HCL pulsatile drug delivery system. Using the quadratic polynomial model to establish two objective functions, the two adjusted coefficient of determinations(Ra2) are 0.983 and 0.975, the model imitation have a good result. We get a series of Pareto non-inferior solutions that are optimized by non-dominated sorting genetic algorithm-II. Solution 14th: when the proportion of ethyl cellulose in the coating is 80.82% and the coating weight level reaches 6.61%, the drug release lag time is 211.78min; the cumulative release is 92.77%. Solution 29th: when the proportion of ethyl cellulose in the coating is 80.70% and the coating weight level reaches 8.10%, the drug release lag time is 220.78min; the cumulative release is 90.23%. The original text use the Response Surface Method(RSM) to optimize the objective functions, the result is that when the proportion of ethyl cellulose in the coating is 78.75% and the coating weight level reaches 7%, the drug release lag time is 203.51min; the cumulative release is 95.57%. Although, the cumulative release that optimized by non-dominated sorting genetic algorithm-II is lower than the original text’s result, the drug release lag time in original text doesn’t keep within the recommended limits. The two solutions in this article optimized by non-dominated sorting genetic algorithm-II are kept in the recommended limits, and the cumulative release is all higher than 90%. There are 24 solutions’ lag times within the specified range and their cumulative releases are higher than 80% in 30 random search solutions. In the optimization study of Zaleplon pulsatile release micro pill, Using the quadratic polynomial model to establish two objective functions, the two adjusted coefficient of determinations(Ra2) are 0.946 and 0.877, the model imitation have a good result. The relatively optimal solution that optimized by non-dominated sorting genetic algorithm-II is that when the swelling layer coating weight increase 22.38%, the weight ratio between CC-NA and HPMC is 6.94:1, and the controlled release layer coating increase 35.44%, the drug cumulative release of 4h is 9.02%, and the cumulative release of 6h is 84.28%. The result of the original text optimized by RSM is that when the swelling layer coating weight increase 26.00%, the weight ratio between CC-NA and HPMC is 7:1, and the controlled release layer coating increase 36.64%, the drug cumulative release of 4h is 6.54%, and the cumulative release of 6h is 83.71%. Compared with the result of the original, the drug cumulative release of 4h is 9.02% less than 10%, this guarantee that the drug release reaches the specified lag time. And the drug cumulative release of 6h is 84.28% higher than 83.71%, the drug release is more complete. In 30 random search solutions, there are 23 solutions within the cumulative release of 4h and 6h range at the same time. Conclusion:There will be subjective when choosing the optimal solution using the Response Surface Method(RSM), and when the factors are more than 4, this method isn’t able to obtain the response surface. So, RSM has some drawbacks. Using the non-dominated sorting genetic algorithm-II to optimize the objective functions that have been transformed, the results are satisfactory. This algorithm is scientific and the optimization program is reliable, it can also provide a Pareto optimal set. NSGA-II could be considered as a more scientific statistical method to optimize the formulation of pulsatile drug delivery system, and providing a new way to solve the problem that the optimization goal is a range not a point. |
PDF Full Text Request