Abstract: In order to avoid the problem that drilling cannot continue to work due to material shortage, warehouses often distribute pipelines and casings more than planed demand in the production of oilfield. In some cases, the excess pipelines and casings can't be used up. And if unused pipelines and casings are not recycled effectively, resources would be wasted. But the recycling would be in vain if the recycling cost is higher than the value of pipelines and casings. Based on the above, the Vehicle Routing Problem considering recycling value of pipelines and casings is studied, and a model which has a special constraint which requires that the recovery costs are lower than the current value of the pipelines and casings is proposed, and improved Differential Evolution Algorithm which is more effective than Genetic Algorithm is implemented. Numerical experiments are performed on real data from the south oilfield of the Ordos Basin belonging to North China branch of Sinope. The recovery scheme of the 94 oil wells includes 17 paths, and the transportation cost is RMB 37,200 lower than the value of oil pipelines and oil casings which is RMB 150,300. Finally, the sensitivity analyses performed examine the robustness of the proposed model. The results show that the unit price directly affects the recovery expense, and the number of vehicles is the most affected by the amount of unused pipelines and casings. In particular, our study not only has practical significance for saving resources and reducing the cost in the oil-field industry, but also has a great theoretical significance for the theoretical research of reverse logistics.

Key words: oil-field, vehicle routing problem, differential evolution algorithm, niching clearing, sensitivity analysis