Answers are grounded in this paper and cite their sources.
Artificial intelligence offers new opportunities to improve biotechnology engineering by converting complex process data into predictive and strategically useful information. This study developed an AI-driven decision-support framework integrating computational intelligence, explainable modelling, and multi-objective strategic analysis. Biomass concentration, substrate concentration, dissolved oxygen, process temperature, energy consumption, and operating cost were evaluated as predictors of product yield. Six modelling approaches were compared using R², RMSE, MAE, and MAPE. The proposed hybrid AI model achieved the strongest predictive performance, with an R² of 0.967, RMSE of 2.18, MAE of 1.64, and MAPE of 2.09%. SHAP analysis identified biomass concentration (27.8%), dissolved oxygen (21.6%), and substrate concentration (18.9%) as the most influential predictors. Multi-objective analysis demonstrated that maximizing productivity alone increased economic and energy requirements, whereas the AI-balanced strategy provided a more favorable compromise. This strategy achieved a predicted yield of 90.2%, reduced operating cost by 10.2% and energy demand by 20.3%, and improved resource efficiency from 71.2% to 87.4% compared with conventional operation. The findings demonstrate how predictive accuracy, model interpretability, and strategic optimization can be integrated to support balanced biotechnology engineering decisions involving technical, economic, resource, and operational considerations.
No takes yet. Share an insight, caveat, or question.
Saxena et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: