That waste generation, implementation of the home appliance trade policy and opportunities of free returns from shopping online, are promoting the development of reverse logistics network design. Currently, researches of the reverse logistics network are concentrated in remanufacturing and reusing. Because remanufacture can make high profits, while the process of reusing is very simple. But a little research is about recycling. Moreover, aimed at different types of recycled products, joint research of the three reverse logistics network is rare. An enterprise is impregnable in the market economy full of fierce competition only when it has considered the recycling of product. Therefore, this study makes an important theoretical and practical sense on the development of enterprise. Then, an overall introduction of the three reverse networks is given based on the analysis of recycled products’ characteristics. Major work and results are as follows:(1) A game model among OEM, retailers,3PL and social is built. Different recovery decision-makers are chosen for different reverse logistics networks. Multiple stakeholders involving in the reverse logistics network were weighed in the multi-objective mixed integer programming model. Meanwhile, a bi-objective particle swarm algorithm and two-stage fuzzy constraint programming method are used in the model.(2) In response to the state mandatory recycling of waste materials strategy, the economic costs of reverse logistics network model takes government subsidies factors into account. For the re-manufactured and re-used products, it considered government subsidies related to recovery rate; for recycled products, it weighed government subsidies relevant to the type and amount of recycled products in view of raw materials for different products.(3) The dynamic location model of reverse logistics network is developed in the uncertain environment. When the time span is large, customer demand would be changed obviously. Therefore, the amount of waste will change, and the corresponding recovery outlets will also be affected. To this end, a multi-period dynamic reverse logistics network model is established. At the same time, for the amount recovery under uncertainty, the reverse logistics network model is constructed, aimed at the random and fuzzy environment. |