Research Article
Explainable Machine Learning Framework for Sustainable Supply Chain and Operational Decision Making
@INPROCEEDINGS{10.4108/eai.28-4-2025.2358026, author={Jagdish Nathumal Utwani and Deep Mangat and Bharadwaja K and Tantepudi Laxmi Poojitha and Prasad Babu Jayanthi and Vidya Sagar S. D}, title={Explainable Machine Learning Framework for Sustainable Supply Chain and Operational Decision Making}, proceedings={Proceedings of the 4th International Conference on Information Technology, Civil Innovation, Science, and Management, ICITSM 2025, 28-29 April 2025, Tiruchengode, Tamil Nadu, India, Part II}, publisher={EAI}, proceedings_a={ICITSM PART II}, year={2025}, month={10}, keywords={sustainable supply chain; explainable artificial intelligence (xai); multi-objective optimization; machine learning; lifecycle emissions estimation; decision support systems}, doi={10.4108/eai.28-4-2025.2358026} }- Jagdish Nathumal Utwani
Deep Mangat
Bharadwaja K
Tantepudi Laxmi Poojitha
Prasad Babu Jayanthi
Vidya Sagar S. D
Year: 2025
Explainable Machine Learning Framework for Sustainable Supply Chain and Operational Decision Making
ICITSM PART II
EAI
DOI: 10.4108/eai.28-4-2025.2358026
Abstract
With a growing number of complex and sustainability driven supply chains, in general, traditional cost decision making does not consider tradeoffs between cost, environmental impact, and operational efficiency. An explainable machine learning based decision support framework integrating predictive analytics, lifecycle emissions estimation and multi objective optimization is proposed that guides selection of supplier, transport mode and conducting operational scheduling. It then applies XGBoost models for predictions on cost and emissions, NSGA-II with Pareto optimization and SHAP with counterfactual analysis to provide interpretable recommendations on the custom generated synthetic dataset for real world logistics and production parameters. Results are then evaluated comparatively and show a 12% in cost, 32% in emission, 3-day reduction in lead time, improvement in reliability and sustainability scores. The inclusion of explainable AI for increasing the transparency and trust in this system makes it practically adoptable for the real world. This work narrows the gap of data–driven optimization and the sustainable and transparent supply chain decision making.
