A new data-driven robust optimization method for sustainable waste-to-energy supply chain network design problem

In this study, the efficient management about waste transportation and energy production is addressed. For this purpose, a new sustainable waste-to-energy (SWTE) supply chain network is designed by considering the facility adjustment strategy. When sufficient historical data about uncertain transportation costs are available, data-driven uncertainty sets are constructed by the support vector clustering (SVC) technique. Based on the constructed uncertainty sets, a novel SVC-based data-driven robust optimization (RO) model is developed for our SWTE problem. Since the proposed model is a nonlinear semi-infinite programming model, we adopt linearization and duality methods to reformulate it as its equivalent mixed-integer linear programming (MILP) model. From the numerical experiments on a practical case, we find that our new model can achieve a significant economic increase of approximately 2.30% compared to the budgeted RO model. Furthermore, several useful managerial implications are obtained from the computational results: (i) The proposed RO method can provide robust rather than conservative optimal decisions; (ii) adjusting regularization parameters can strike a balance between robustness and conservativeness in optimal decision-making; and (iii) insufficient capital budget or excessively high sustainability threshold will result in revenue losses.