Robust location of collection facilities for waste management

This paper presents a robust optimization model for the location of collection facilities in municipal waste management under uncertain demand. A two-tier collection system is considered where waste generated by households and businesses is aggregated in local transshipment facilities and then transported by mother vehicles to final treatment sites. The problem is modeled as a capacitated facility location problem and extended using the Bertsimas–Sim robust optimization framework to hedge against fluctuations in waste demand explicitly. A deterministic mixed-integer linear programming model is first formulated and next its robust counterpart is derived by introducing a budget-of-uncertainty parameter that controls the trade-off between cost and protection level. The model is applied to a real case study in Latina (Italy), using detailed geo-referenced data on users, candidate transshipment points, and available vehicles. Computational experiments show that the robust model can reduce worst-case total costs by up to several tens of percent compared to a purely deterministic design, at the price of a moderate increase in nominal operating costs. The results demonstrate that robust facility configuration tends to open a slightly larger number of strategically located transshipment points, resulting in shorter average collection distances and improved resilience to demand surges. Managerial implications for local authorities and waste management companies are discussed, and promising directions for extending the model to multi-period and multi-waste-stream planning are identified.