Uncertainty Model And Collection Forecast In Reverse Logistics

There're many reasons which lead more attention to the Reverse Logistics(RL for short), including the environmental problems about the breakage of ecological balance, resource exhaust, huge living or industrial junk, and the development of high-tech, also with the need to increase the competitive power for the company. Though it has made great progress over the past decade, there's still a tough problem of uncertainty which limits the improvement of these theories and practices. The thesis is about the analysis and resolution of the problem. There're five parts in the thesis.The first part is the review of the RL. The background and meaning of the thesis are pointed out. The definition of closed-loop supply chain and RL, as well as the relationship between them are introduced. Finally, the uncertainty problem is put forward on the basis of the research status of RL domestic and abroad, especially the research on model.The second part is mainly about the factors of the RL uncertainty and an analysis model. At first, the thesis makes a contrast about four typical markets by process; and then based of it, the thesis summarizes the direct factors of uncertainty and explores the essential factor corresponding to the direct factors. The thesis also puts an analysis model by AHP to analysis these factors. The specialized high-tech market and the daily consumer market are picked up to show the calculate processes.The focus of the third part is the resolution of the uncertainty problem. The relative researches before are summed up. The principles of forecast technologies are introduced, especially the time series forecast and grey forecast. The thesis pointed out the application of forecast to resolve the quantity uncertainty of collection process.The forth part sets an example to show the calculation process by time series forecast and grey forecast, and analyzes the simulation degree by excel. It comes to the conclusion that the grey model is more suitable for the RL with high degree of uncertainty.In the last part, the main results and the limitations of the method are generalized.