ABSTRACT
To mitigate the spread of novel coronavirus, how to optimise COVID-19 medical waste location-transport strategies remains an open but urgent issue. In this paper, a novel digital twin-driven conceptual framework is proposed to improve the strategic decision on the location of temporary disposal centres and, subsequently, the operational decision on the transport of COVID-19 medical waste in the presence of hierarchical relationships amongst stakeholders, circular economy, uncertainty in infection probability, and service level. The circular economy aspect is measured by the reduction of infection risks and costs, as well as limiting exhaust emissions. The polyhedral uncertainty set is introduced to characterise stochastic infection probability. Digital twin technology is further used to estimate the upper and lower bound of the uncertainty set. Such a problem is formulated as a digital twin-driven robust bi-level mixed-integer programming model to minimise total infection risks on the upper level and total costs on the lower level. A hybrid solution strategy is designed to combine dual theory, Karush-Kuhn-Tucker (KKT) conditions, and a branch-and-bound approach. Finally, a real case study from Maharashtra in India is presented to evaluate the proposed model. Results demonstrate that the solution strategy performs well for such a complex problem because the CPU time required to conduct all experiments is less than one hour. Under a given uncertainty level of 36 and perturbation ratio of 20%, a regional transport strategy is preferred from generation points to transfer points, while a cross-regional one is usually implemented from transfer points to disposal centres. It is of significance to determine the bound of available temporary disposal centres. Using digital technology (e.g., digital twin) to accurately estimate the amount of COVID-19 medical waste is beneficial for controlling the pandemic. Reducing infection risks relative to cost is the prioritised goal in cleaning up COVID-19 medical waste within a relatively long period.
Fuente: Computers & Industrial Engineering
Available online 20 February 2023
