Predictive modeling and multi-objective optimization of 3D-printed PLA parts using a hybrid random vector functional link (RVFL) and henry gas solubility optimization (HGSO) framework

To forecast the mechanical characteristics of 3D printed Polylactic Acid (PLA) specimens, this work explores artificial intelligence models. Tensile strength (TS) and surface roughness (Ra) of printed parts were examined in relation to five process parameters: build angle (BA), infill angle (IA), layer thickness (LT), printing speed (PS), and nozzle temperature (NT). In the machine learning modelling stage models such as Random Vector Functional Link (RVFL), Random Vector Functional Link tuned by Henry Gas Solubility Optimization (RVFL-HGSO), and Artificial Neural Network (ANN) were created to forecast the printed parts’ surface roughness and tensile strength. Moreover this study shows the effectiveness of random vector functional link model tuned by the Henry Gas Solubility Optimization (RVFL-HGSO) when predicting surface roughness. The developed model achieved highest R2 = 98.95%, RMSE = 0.777, MAPE = 1.002%, and MAE = 0.514 on test dataset. It increased R2 by 4.98% when compared with RVFL and by 9.93% when compared with ANN on test dataset. RVFL-HGSO decreased MAPE by 49.93%, also MAE by 53.90%, and RMSE by 29.68% when compared with RVFL on test dataset. It also, decreased MAPE by 69.02%, also MAE by 73.95%, and RMSE by 58.49% when compared with ANN. The multi-objective optimization by particle swarm optimization (MOPSO) gives best possible solution at place of BA = 44.16, NT = 220, IA = 14.93, LT = 0.22, PS = 45.18: Minimum predicted Ra = 2.53 and maximum predicted TS = 44.15. The findings demonstrate the RVFL-HGSO model’s unique method for finding intricate correlations in the dataset, laying the groundwork for better product design and mechanical property optimization through process parameter adjustment.