تعیین اندازه دسته تولید در برنامه ریزی نیازمندیهای مواد معکوس برای جداسازی قطعات برنامه ریزی

Gupta and Taleb (1994. Scheduling disassembly. International Journal of Production Research,32(8), 1857–1866) presented an algorithm for reverse MRP that can be applied to a product structure in which there is a certain demand for components and a need to know the number of products to disassemble in order to fulfil the demand for those components. However, that algorithm did not consider lot sizing at all. Incorporating lot sizing makes the problem a great deal more challenging in disassembly situations because of the disparity between the number of components present in the product and the demand of the components. This is because the demand for the components is often not in the same proportion as their existence in the product structure causing excess inventories for the components with low demand compared to their brother items. In this paper we present an algorithm for reverse MRP to facilitate the use of lot sizing. An example is presented and its implementation is analysed.

MRP (materials requirements planning) is a well established and a widely used production planning procedure. However, MRP is an assembly oriented scheduling system, which cannot be applied to shop floor operations for disassembly scheduling (DS). Panisset (1988) also pointed out that “most MRP logic (and the supporting bill of materials) do not provide facilities to plan disassembly”. However, partial disassembly and reassembly for maintaining and repairing complex products have been practiced for a long time. Often, discarded products are disassembled for parts and materials for repairing and remanufacturing products (Lambert and Gupta, 2005). The procedure to accomplish DS is known as reverse materials requirements planning (RMRP), since the procedure is basically a reversed form of the regular MRP (Gupta and Taleb, 1994). There is an interesting difference between the calculations for a regular MRP tableau and reversing it. A very important characteristic of regular MRP is that it is used for assembly and there is a single source of demand located at the end item (root). One of the problems that arises in reversing MRP is that we can have multiple sources of demands, viz., the components or the leaf items in the disassembly product structure. Besides, these leaf items are not independent since they all have to be fulfilled by the same root item from which they originate (Gupta and Taleb, 1994). This condition adds a layer of complexity to the calculations as it produces excess inventories for the components with a relatively low demand compared to their brother items (Gupta and Taleb, 1994). This happens because often, the demand for the components is not in the same proportion as the proportion in which they are present in the disassembly product structure. While several researchers have made use of RMRP (see, for example Inderfurth and Langella, 2006; Kim et al., 2006; Lee and Xirouchakis, 2004), no one has addressed the concept of lot sizing in connection with DS. In this paper, we present an algorithm for RMRP to facilitate the use of lot sizing. This paper is organized as follows. A brief literature review is provided in Section 2. The proposed algorithm is described in Section 3. An example is also included to illustrate the implementation of the algorithm. Section 4 provides a preliminary experimentation to demonstrate that the use of lot sizing actually reduces the planning costs in reverse systems. Then in Section 5, a simulation study was carried out to test the performances of several lot-sizing rules. The conclusions are presented in Section 6.

This paper incorporated lot sizing in RMRP. To that end, we proposed an algorithm to facilitate its use. The proposed algorithm complements the original Gupta and Taleb (1994) algorithm and can be applied to real life problems, as it is easy to implement on a computer and have useful applications in the planning of disassembly of products to fulfil the leaf items’ demands. In addition, this algorithm could encourage companies to develop marketing strategies to reduce the prices of leaf items and thus encourage other companies to find new applications for them. To demonstrate the economic effectiveness of using lot sizing, we performed an experimentation analysing products with various structures, several cost scenarios, and three lot-sizing techniques used at different levels of the BOM. The results using ANOVA revealed that the lot-sizing rules, POQ and EOQ, are indeed cost effective at upper levels of the structure independent of the complexity of the structure, while LFL is a better option at the lowest level, especially in simple structures. On the other hand, the lot-sizing techniques used at the intermediate levels of the structure or in ordering the final product do not seem to play an important role. This research can be extended in several ways. First, there is a need to develop an algorithm to automatically obtain optimal solutions in DS of the root. Second, a modification of the algorithm is needed when it is necessary to accommodate parts and materials commonality. Finally, it would be of interest to relax the deterministic assumptions and extend the methodology to operate in uncertain or stochastic environment so as to mimic real life scenarios.