人工神经网络多指标同步优化阿替洛尔非pH依赖型缓释片

Application of an artificial neural network in the optimization of atenolol pH-independent sustained release dosage forms

目的:应用人工神经网络筛选阿替洛尔非pH依赖型缓释片的处方。方法:以羟丙甲纤维素为缓释材料,琥珀酸为pH缓冲剂研制非pH依赖型阿替洛尔缓释片,以不同用量的羟丙甲纤维素和琥珀酸配制成不同处方缓释片,测定各处方在pH7.4和0.1mol/LHCl中的释放度并计算得分,将此数据用于训练BP人工神经网络,并将此网络用于处方筛选。结果:优化处方在不同pH介质中均有较好的缓释效果。结论:人工神经网络多指标同步优化法可用于处方筛选。

Objective: To optimize atenolol pH-independent sustained-release tablets using artificial neural network (ANN) . Methods: Atenolol pH-independent sustained-release tablets were prepared by using hydroxypropyl methyllose(HPMC) and succinic acid as excipients. The various contents of HPMC and succinic acid for 20 tablet formulations were used as ANN model input. In vitro accumulation released in pH7.4 and 0.1mol/L HCl solutions was determined and the marks were computed. The values were used as output and the constructed ANN model was used to optimize the tablet formulations based on desired target in vitro release profiles. Results: The prepared tablets possessed good sustained-releases in different solutions. Conclusion: ANN can be used to design the sustained-release tablet formulation.