Achieving Wealth from Bio-Waste in a Nationwide Supply Chain Setup under Uncertain Environment Through Data Driven Robust Optimization Approach

Addressing dual crisis of fossil fuels i.e. environmental as well as gradually decreasing reserves, design of a nationwide robust supply chain network based on bio-energy has been presented in this work. The four echelon supply chain caters the target of blending 20% bio fuels with conventional fuel during 2018–2026. The mixed integer linear programming model considers handling of multiple types of feed sources, products, transport options while performing the techno, economic and environmental analysis of the supply chain to optimally determine the operational and design decisions. The uncertainty in demand, import product price and biomass feed supply has been tackled using data driven robust optimization approach developed using fuzzy transcription of discontinuous uncertain parameter space using machine learning based unsupervised learning methods. To handle a ∼50% increase in overall biofuel demand over the nine-year planning horizon, the deterministic model shows a dynamically changing supply chain with a ∼20% increase in the newly added locations; however, the worst case robust optimization results are reported to be 6% leaner than the results obtained for the deterministic model. The sensitivity analysis of biomass availability on net present value indicates the need of 43% and above biomass feed supply to run such bio supply chain sector to survive.