طراحی سیستم کنترل موجودی با ادغام طبقه بندی موجودی و انتخاب سیاست

Very large numbers of inventory items complicate the inventory control process. Companies categorize their inventory items into a few groups and take similar inventory control policies for the items in each group to overcome this problem. In this regard many grouping methods have been proposed. Some researchers have studied the appropriate inventory policy for each group. Since both the actions of categorization and policy selection are sub-optimal solutions for the original problem of efficient inventory control policy, this paper proposes an integrated model to categorize the items and find the best policy simultaneously. As it is difficult to find a global solution, simulated annealing is used to find appropriate solutions. The model results are compared with the findings of other methods both for dissimilarity and total inventory values.

Classification is widely proposed in the literature to tackle the problem of very large numbers of inventory items. Thousands of inventory items in companies even with moderate size increase the risk of losing sight of the most important items and spending unnecessary resources in controlling less important ones. Therefore companies try to classify the items and select appropriate control policies for each group. In this regard many authors have studied the classification process and proposed various exact and heuristic methods to classify inventories satisfying some criteria. ABC classification is the most widely employed technique, which in its basic form considers the only criteria of annual use value (Cohen and Ernst, 1988). Some other authors have focused on the appropriate control policies for each group of the items. Reorder point, two-bin systems and material requirement planning (MRP) are some of the developed strategies (Hautaniemi and Pirttila, 1999). Thus, in the literature, the original problem of making an effective inventory control system has been decomposed into two problems of classification of inventory items and finding appropriate strategies for each group. This has misled some authors to focus on decomposed sub-problems independently and forget the original goal. Although this may produce some sub-optimal solution for the original problem we should not forget that the aim of the classification is not solely to classify items but to excel in the performance of inventory control policy. In this paper we propose a model that concurrently classifies inventory items and selects appropriate policies for each product group with the objective of having an effective inventory performance. The paper is organized as follows. In Section 2 the related literature is reviewed. The developed model is described in Section 3. The coding of the problem in simulated annealing is described in Section 4. An illustrative example is solved in Section 5. Section 6 discusses the algorithm's time complexity and finally Section 7 presents the conclusions.

In the literature, the original goal of having an efficient inventory control policy has been of less concern in many papers than concentrating on sub-optimal problems of inventory items categorization and policy selection. In this paper an integrated method is proposed to simultaneously categorize inventory items and find an efficient control policy. The proposed model is compared with annual dollar usage, AHP weighted score, the method proposed by Zhang et al. (2001) and optimal inventory score, and exceeds all of them in minimizing both dissimilarity and total inventory value. Simulated annealing is used for the model to find a good solution. It is shown that the run time of the simulated annealing is a polynomial of order two with respect to sample size.