La Reverse Logistics nell'industria manifatturiera: l'impatto dei processi di recovery sulla formulazione dei piani di produzione e di approvvigionamento

There are several ways to design research and choose methods of data collection, considering the selection of the research topic, the research paradigms, the research questions, etc. According to the preeminent literature, the scientific method is a procedure by which the visibility in the "unknown" is achieved through a cyclical process, which needs to be addressed in the following phases 1) PROBLEM ANALYSIS 2) THEORETICAL FRAMEWORK AND KNOWING STATE 3) RESEARCH OBJECTIVES 4) TECHNICAL ANALYSIS 5) HYPOTHESIS AND RESULTS  PROBLEM ANALYSIS: It can be seen from the analysis of the Reverse Logistics reference literature and in particular in the Production Planning field, there are few studies that investigate the problem of the EOL impact in terms of costs within an MRP process. Among the most interesting studies we can point out: Ferrer & Whybark (2001) have developed an MRP-based methodology able to give to the users answers to questions like: how many and which components to buy, what mix of components to disassemble, and what kind of components need to de assembled in order to meet and satisfy the demand. Souza & Ketzenberg (2002) and Souza & al. (2002) have used a "two-stage GI / G / 1 queue-network" model in order to identify the Long Term product mix able to maximize the profit of the companies that wish to meet the demand of refurbished products, the demand of only new products or a mix of both. The authors used a DES model (Discrete Event Simulation) which takes in regard some real parameters such as stochastic returns trends and stochastic return production, that are essential to test the robustness of the model. Jayataman (2006), through a mathematical programming model, has determined the optimum number of components, having a certain level of quality, which must be acquired yearly, disassembled, refurbished and stored in a given period of time; Denizel et al. (2010) have developed a multi-period planning model that considered the uncertainty of the EOL products quality.  THEORETICAL FRAMEWORK AND KNOWING STATE: In the field of Reverse Logistics is hard to find a complete definition of the RL. Analyzing of the reference literature, we can note that over the years there has been a profound evolution of matter, which has grown from a simple study of the goods "return flows" from the consumer or another link in the distribution chain up to the OEM until to become, starting from 1999, a real " planning process, implementation and control of the flows suitable to generate value to the company. The Reverse Logistics can be defined as a "genus", which deploys its ramifications in different fields. In fact, in the literature there are many studies that range from the Inventory Management to the study of operational issues that involve the operative production planning. It also points out that in recent years, numerous studies have flourished to put order in the matter, and in this sense we reference: SMGupta "Reverse Supply Chains- Issues and Analysis" CRC Press, 2013; MAIlgin-SMGupta "Remanufacturing Modeling and Analysis" CRC Press, 2012; KKPochampally-S.Nukala-SMGupta "Strategic Planning models for Reverse and Closed-Loop Supply Chain" CRC Press, 2009; K.Govindan-H.SoleimaniD.Kannan "Reverse Logistics and closed-loop supply chain: a comprehensive review to explore the future," European Journal of Operational Research 240 (2015) 603-628; C.NussR.Sahamie-D.Stindt "The Reverse Supply Chain Planning Matrix: A classification scheme for planning problems in reverse logistics," International Journal of Management Reviews (2014).  RESEARCH OBJECTIVES: Taking advantage of the studies conducted during the three years of the doctorate, we have focused our attention on how the management of returns and non-compliant products during the production process impacts on the MRP cost. The realization of the MRP model finds its roots in the fact that, as Gupta (1994) has theorized, there are many difficulties in planning the stocks due to the uncertainty related to the impossibility of predicting the rate of raw material recovered from returned products; Impossibility to know a priori the conditions of the product as long as this is not inspected; Need to disassemble the product; Difficulties in demand management related to the difficulties in balancing between returned products offer and relative product demand: Difficulties in the flows coordination of materials resulting from the dismantling of products with peculiar characteristics and that can be reused only in the reassembly of certain products; Uncertainty related to the returned products timing and amount. These issues have an impact on the determination of the MRP total cost, but these effects could be annulled by the savings in virgin raw materials, purchasing through appropriate disassembly processes. • TECHNICAL ANALYSIS: In order to demonstrate what was held in the research objectives, it was created a discrete event simulation model (DES) through the simulator Arena. In this model we wanted to describe how a product, having a certain residual quality, undergoes the process of recovery and what are the costs and time associated to the performing of this operation. The recovered materials are then inserted within the MRP, in which it is deduced on the POH and, consequently, the ORP is affected by the presence of these products recovered. • HYPOTHESIS AND RESULTS In order to analyze the convenience of a MRPR system, it was decided to analyze the case where the BOM of the product was structured according to multiple levels and in the case where the product bill of materials is monolevel (flat). Then, within the simulation software Arena, different scenarios has been hypothesized by varying the following variables: Recovery Plan, NC Plan; Quality Recovery; Disassembly Mean Time; Disassembly Time Deviation. They haven’t been changed the NC product quality level and the time of inspection, because, in the evidence of the company experience, was proved: that the quality of the NC uncomfortable product is always high because they wasn’t never used and disassembled in a short time while , regarding the inspection time, the company evidences show that this corporate timescale can be well described by an uniform distribution ranging between 4 and 8 minutes. Thanks to the ANOVA analysis, it is demonstrate that the role of the BOM in the process of disassembly is preeminent, in particular, the interaction between the BOM and the flow of Returns, between the BOM and the Flow of NonConformity, and between the BOM and time Disassembly is particularly significant (p <2e16). So it can be said that what was held in the Box Plot earlier, that describes how the relationship between the BOM and the Middle Disassembly Time that affects the savings in terms of procurement cost is not a coincidence. Furthermore continuing scanning the table, we can get another important fact: the interaction between Middle Disassembly Time and the recovery plan is particularly significant in determining the savings in terms of cost within an MRP plan (p = 1.77e -12), while the mutual influence that occurs between the recovery plan of the non-compliant products and Middle Disassembly Time appears to be significant but are less important. The thesis has four chapters. In the first, starting from the concept of Closed Loop Supply Chain, the attention is focused on the Reverse Logistics, identifying essential institutions and stakeholders involved. A particular attention is given to the product recovery concept, which will be the workhorse of the chapter related to the MRP model. The second chapter presents a review of Reverse Logistics literature, focused on strategic issues, on the problem concerning the Planning & Control and the Inventory Management. The chapter concludes with a review of the literature of the items considered most important published until 2013. The third chapter describes the MRP model, starting from the definition of the problem until its mathematical formulation. The data collected will then be analyzed in the fourth chapter, through the use of common statistical inference techniques. [edited by Author]