A two-stage interval-stochastic programming model for planning end-of-life vehicles allocation under uncertainty

• A two-stage interval-stochastic programming model for planning ELVs allocation is developed.
• The model can directly handle uncertainties expressed as either PDFs or intervals.
• A hypothetical case study demonstrated the potentials of the formulated model.
• The model can identify ELV allocation targets that will secure maximized profits and minimized disruption risks.
• The model is applicable across vehicle recycling industry.

The ultimate fate of vehicles can vary, including being recycled, exported, abandoned, illegally treated in unauthorized facilities, stolen and processed for parts or illegally exported, etc. Nowadays, export of used vehicles represents the most significant barrier to the more efficient vehicle recycling in the EU, because millions of vehicles (mainly from Germany, Italy, UK and France), which are expected to go to domestic vehicle recycling factories are exported. As a result, how to allocate limited and frequently insufficient quantities of collected end-of-life vehicles (ELVs) to satisfy vast demands of vehicle recycling factories becomes a significant concern of many government or private authorities that control the ELV collection and treatment networks across the EU. In this paper, a two-stage interval-stochastic programming model is developed for supporting the management of ELV allocation under uncertainty. The formulated model can directly handle uncertainties expressed as either probability density functions or discrete intervals. Moreover, it can support the analysis of various policy scenarios that are associated with different levels of economic penalties when the promised ELV allocation targets are violated. The proposed model has been applied to a hypothetical case study in which several scenarios with different ELV allocation policies were comprehensively analyzed. The obtained results indicated that reasonable solutions had been generated. There is significant influence of parameter uncertainty on model solutions and the analyzed policy scenarios. Various policies in negotiating the ELV allocation targets with vehicle recycling factories would lead to different economic and risk values. Proposed model can serve as the support for authorities to identify ELV allocation targets that will secure maximized profits and minimized disruption risks of the vehicle recycling factories. It is applicable across vehicle recycling industry that processes dozens of millions of ELVs every year.