Optimization of a closed-loop supply chain system considering government incentives mechanism under deep learning algorithms

Against the backdrop of multiple uncertainties, our research endeavors to design a comprehensive closed-loop supply chain system that incorporates dual recycling channels and a series of manufacturing-remanufacturing processes. The focal point of this study lies in the establishment of a holistic profit model aimed at a thorough exploration of the effect of the government incentives mechanism (GIM) on this system. Additionally, we employ deep learning algorithms (DLA), a kind of AI technology, for calculation and solution analysis of the model. The results show: (1) Remanufacturers can develop reasonable recycling, remanufacturing, and manufacturing strategies based on different scenarios; (2) The quality level of recycled products, rather than different demands, has a significant impact on the amount of penalty and reward. Thus, when establishing a GIM to encourage recycling and remanufacturing, the government should primarily focus on the uncertainty of the recycled products’ quality. (3) In the absence of a GIM, remanufacturers are reluctant to strive to improve the recovery rate, and are more inclined to choose informal channels to reduce recovery costs; (4) The GIM can stimulate and regulate enterprises’ recycling activities. Hence, the government should formulate a reasonable GIM to regulate the recycling of enterprises and supervise and guide the transformation and upgrading of informal channels. This research provides profound insights for the establishment of a robust closed-loop supply chain under multiple uncertain environments, and also, this research combines AI technology to improve computational accuracy, provide more reasonable support for business decision-making and government supervision, and provide assistance in realizing a circular economy.