Multi-Objective Spray Drying Process Optimization via BBD-ANN-AGWO Framework: Case of Andrographolide Amorphous Solid Dispersions.

This study aims to establish a multi-objective spray drying process optimization framework, with andrographolide (ADG) amorphous solid dispersion serving as a model drug. The proposed framework integrates Box-Behnken design (BBD), artificial neural networks (ANN), and an adaptive grey wolf optimizer (AGWO). Molecular docking was employed to screen potential polymeric carriers, and in vitro dissolution experiments were conducted to identify the carrier with the best solubilizing performance. Single-factor experiment and BBD trial were carried out to evaluate the effects of key process parameters on energy consumption, drying loss, drug release, and yield. A multi-response ANN model was then developed based on BBD data. To enhance the model's generalization ability and account for process variability, virtual samples were introduced to augment the training dataset. AGWO was subsequently applied to perform inverse optimization and identify the optimal process parameter combination. Experimental validation demonstrated good agreement with the model predictions. X-ray diffraction analysis (XRD) and Fourier-transform infrared (FT-IR) analyses confirmed the amorphous transformation of ADG and its satisfactory short-term physical stability. The proposed method provides a feasible and intelligent strategy for multi-objective optimization in pharmaceutical formulation development.