تعیین سیاست های مرتب سازی اقلام موجودی در کلاس B با استفاده از پایه قوانین IF-THEN

Multiple criteria ABC inventory classification is one of the most common techniques of planning and controlling of inventory. In this type of classification, other criteria in addition to the criterion of annual dollar usage are taken into account, and then items, based on their priority, are divided into 3 classes with different ordering policies (OPs). In this paper, multiple criteria ABC inventory classification is first performed using the Hadi-Vencheh model (2010). Then suitable OPs are determined for items of classes A and C and Finally, we apply the Zafiropoulos and Dialynas (2005) approach to extract fuzzy rules from the policies of these two classes for determining the policies of items class B. To show applicability of the proposed model, it is then implemented in a soft-drink manufacturing factory.

One of the most important factors affecting the cost of the goods manufactured is determining the suitable policies to order raw materials stock in the warehouse. The duty of the inventory managers is to select a suitable alternative of these policies in such a way that the items required be available to the needed quantity in due time. Each of the different OPs will impose different costs by determining the special quantity of items, lead time, etc. The tight OPs such as fixed order size (FOS), although may impose the high inspection costs, the probability facing stock-out will be reduced. Conversely, the inexact OPs as twin bin (TB), despite the lower inspection costs, will have the higher stock-out probability. Therefore, an able inventory manager should adopt best alternative to determine an efficient policy, based on the priority each item. To achieve such a target, the classifying items inside the groups with different priorities should first be acted and then, suitable policies should be adopted for each group (Chen, Li, Kilgour, & Hipel, 2008). ABC classification is one of the most common techniques of classification, dividing items into 3 classes, namely, A (very important), B (moderately) and C (least important), based on Pareto principle. Traditional ABC (TABC) only uses the criterion of annual dollar usage, but many papers have mentioned that in addition to this criteria, other criteria as ordering cost, criticality of part, lead time, commonality, obsolescence, repairability, number of requests, scarcity, durability, perishability, reparability, demand distribution, stockability (Cakir and Canbolat, 2008, Chen et al., 2008, Flores and Whybark, 1986, Flores and Whybark, 1987, Guvenir and Erel, 1998, Hadi-Vencheh, 2010, Ng, 2007, Partovi and Anandarajan, 2002 and Ramanathan, 2006) are also needed for classification. Many approaches have been proposed for multiple criteria ABC inventory classification, including the bi-criteria matrix approach (Flores and Whybark, 1986 and Flores and Whybark, 1987), a case-based distance model (Chen et al., 2008), a artificial neural network approach (Partovi & Anandarajan, 2002), a weighted linear optimization model (Ramanathan, 2006), a cluster analysis method (Cohen & Ernst, 1988), a genetic algorithm approach (Guvenir & Erel, 1998), AHP (FAHP) method (Cakir and Canbolat, 2008, Partovi and Burton, 1993 and Partovi and Hopton, 1994). But the available papers on the determination of type of policies of class B are few, and most of them have failed to present a certain suitable policy for the items this class, which is between the classes A and C. However, Silver, Pyke, and Peterson (1998) have suggested several policies on this class, and some articles as Cakir and Canbolat (2008) on the items in class B has introduced both FOS and fixed order interval (FOI). But inventory managers are still confused in adopting policies on this class, since either no discussion has been held over determining the certain OPs of class B, or this issue has been discussed with ambiguity. In this paper, the certain suitable ordering systems for the items of classes A and C are first presented, and then fuzzy rules are extracted for such policies using the Zafiropoulos and Dialynas approach (2005) and Khanlari, Mohammadi, and Sohrabi (2008). Next, by means of these rules, the suitable policies are presented for the items of class B as well. Framework of this article is as follows. Section 2 summarizes concepts of fuzzy sets theory, fuzzy number and fuzzy linguistic variable. In Section 3, the suggested model for multiple criteria ABC inventory classification as well as for determining OPs of the items in class B is offered. An illustrative case study is presented to implement the proposed model in Section 4 and finally, in Section 5, conclusion and limitations are offered.

In this paper, we could respond to the uncertainty about determination of policies of items in class B. In most of the articles relating to the ABC classification, determination of OPs of this class has always been ambiguous, and none of them could be ever presented suitable policies for items this class. Thus, we applied fuzzy rules extracted “If–Then” in classes A and C (by means of Fig. 2, Fig. 3, Fig. 4, Fig. 5 and Fig. 6). To extract fuzzy rules, 5 items from class A and 6 items from class C were only used. Since there are many items in class C, an attempt was made to select the items of this class in such a way that usable fuzzy rules are generated to determine the OPs of items B. This model systematically formulated the policies of items in class B. By implementing this model in a soft-drink manufacture factory and submitting these policies to experts, they also verified reliability of such policies. Likewise, in an interview with the experts for the items in class A, FOS ordering system, and for items C, FOI and TB ordering policies were adopted. Nevertheless, in other industries and based on the available criteria relating to that industry, such other policies such as JIT, barcode and automated record can be used for item in class A and VR policy for class C. Model of determining the policies of items in class B has some limitations too as follows: 1. Time and budget spent on collecting data, extracting fuzzy rules and in general, number of its stages is very much, especially when the number of criteria increase, extraction of fuzzy rules is faced with problem, consuming much time. 2. The problems relating to subjectivity of information obtained from experts may bring about contradictory information and rules. 3. Choosing suitable items from classes A and C to extract fuzzy rules is one of the limitations of this model. If unsuitable items are selected from classes A and C, fuzzy rules will not only be unusable, but also they will impose much time and cost on the analysis of the model. Despite these limitations, the model of multi-criteria ABC classification, extraction of fuzzy rules and determining of policies for the items in class B is one of simple, efficient and understandable techniques for managers. In the model presented in this research, any criterion can be easily added and then suitable OPs provided for the items in class B. Thus, its area of implementation for other manufacture industries is usable, in which other different criteria and policies may be used.