Research Information System
Expert Systems with Applications, vol.232, 2023 (SCI-Expanded)
This research develops a multi-stage decision framework for a three-tier collection network for hazardous waste considering random release dates. Applying a cost-based clustering approach, the first stage decisions involve locating the transfer stations and allocating generation nodes to the chosen facilities. The corresponding results, along with a subjective risk-aversion concept and estimated release dates, are utilized to generate an a priori collection plan, which can be further revised once the actual release dates are realized. We performed sensitivity analyses on important model elements, including the subjective risk-aversion parameter and vehicle capacity for collection tours. Our findings indicate that depending on the subjective perspective of the collector, deviations from the original scheduled plan can be a positive or negative phenomenon. For an optimistic decision-maker, adaption to the realized deviations necessitates accepting additional total risk. On the other hand, a pessimistic decision-maker is more likely to benefit from deviations. The findings can provide relevant authorities with practical and realistic strategic and operational decisions.