Delineating the effects of hot-melt extrusion on the performance of a polymeric film using artificial neural networks and an evolutionary algorithm

Abstract

The aim of this study was to utilize an artificial neural network (ANN) in conjunction with an evolutionary algorithm to investigate the relationship between hot melt extrusion (HME) process parameters and vaginal film performance. Investigated HME process parameters were: barrel temperature, screw speed, and feed rate. Investigated film performance attributes were: percent dissolution at 30 min, puncture strength, and drug content. An ANN model was successfully developed and validated with a root mean squared error of 0.043 and 0.098 for training and validation, respectively. Of all three assessed process parameters, the model revealed that barrel temperature has a significant impact on film performance. An increase in barrel temperature resulted in increased dissolution and punctures strength and decreased drug content. Additionally, a successful implementation of an evolutionary algorithm was carried out in order to demonstrate the potential applicability of the developed ANN model in film formulation optimization. In this analysis, the values predicted of film performance attributes were within 1% error of the experimental data. The findings of this study provide a quantitative framework to understand the relationship between HME parameters and film performance. This quantitative framework has the potential to be used for film formulation development and optimization.

Keywords: Artificial neural network; Formulation; Hot melt extrusion; Modeling; Vaginal film.