Quality by design approach for tablet formulations containing spray coated ramipril by using artificial intelligence techniques

Different software programs based on mathematical models have been developed to aid the product development process. Recent developments in mathematics and computer science have resulted in new programs based on artificial neural networks (ANN) techniques. These programs have been used to develop and formulate pharmaceutical products. In this study, intelligent software was used to predict the relationship between the materials that were used in tablet formulation and the tablet specifications and to determine highly detailed information about the interactions between the formulation parameters and the specifications. The input data were generated from historical data and the results obtained from analyzing tablets produced by different formulations. The relative significance of inputs on various outputs such as assay, dissolution in 30 min and crushing strengths, was investigated using the artificial neural networks (ANNs), neurofuzzy logic and genetic programming (FormRules, INForm ANN and GEP). This study indicated that ANN and GEP can be used effectively for optimizing formulations and that GEP can be evaluated statistically because of the openness of its equations. Additionally, FormRules was very helpful for teasing out the relationships between the inputs (formulation variables) and the outputs.