• Machine learning models, particularly random forest, accurately predict the density and mechanical properties of Ti-6Al-4V parts from process parameters.
• Multi-objective optimization using genetic algorithms identified parameter sets that improved tensile strength by 12% and reduced porosity by 15%.
• The approach reduces the need for extensive experimental trials, saving time and material costs in AM process development.
• The methodology can be extended to other materials and AM processes, offering a general framework for data-driven optimization.