ANP

ORIGINAL ARTICLE Make-to-order/make-to-stock partitioning decision using the analytic network process Samira Hemmati & Masoud Rabbani Received: 15 October 2008 /Accepted: 11 September 2009 /Published online: 25 September 2009 # Springer-Verlag London Limited 2009 Abstract This paper presents a decision-making struc- ture to determine the appropriate product delivery strategy for different products in a manufacturing system. These strategies include make-to-stock (MTS), make-to-order (MTO), and hybrid MTS/MTO produc- tion systems. In fact, the proposed approach gets the decision maker the opportunity to benefit from both strategies through applying the hybrid one. There are varieties of driving factors involved in choosing the right product delivery strategy, and all these factors have positive and negative interactions with each others; in this regard, we apply an appropriate multi-criteria decision-making method. In this method, relevant criteria affecting MTS/MTO partition are split into four categories: market-related criteria, product-related crite- ria, process-related criteria, and supplier-related criteria. Due to the interdependency between these criteria, we use analytic network process that generalizes analytic hierarchical process by considering the interdependen- cies among factors. Finally, in order to show the applicability of the proposed structure in practice, the structure is implemented to choose the best production policy among three aforementioned strategies in the real industrial case company. Keywords Make-to-order (MTO) .Make-to-stock (MTS) . Hybrid MTS/MTO .MTS/MTO partitioning . Analytic network process (ANP) 1 Introduction In the face of global competition, a manufacturing company’s survival increasingly depends on how best it can design, manage, and restructure its production system to deal with product diversity, improve delivery reliability, and also reduce system costs [1]. To cope with these issues, manufacturing companies often use different production systems. These production systems can be classified into two major categories: make-to-stock (MTS) and make-to- order (MTO) based on market demands’ response policy [2]. In an MTS system, finished or semi-finished products are manufactured based on anticipating future demands. The main advantage of MTS system is the short delivery time, since the final products are already in stock even before the customer order entry [1]. The main competitive advantage in MTO environments is short delivery time and also the ability to establish reliable delivery dates [2]. Recent years have shown a number of changes in companies’ production policy, and they are gradually moving more to hybrid MTS/MTO production mode. In an MTS/MTO system, a portion of the production system operates as a MTS system and the remaining portion operates in a MTO mode [1]. A proper combination of MTO and MTS can exploit the advantages of both lower inventory and short delivery time. In such systems, some semi-finished products are maintained at one stocking point, so the production delay is just the time needed for the MTO stage. In the related literature, this stocking point is called order penetration point (OPP) in the production line. Generally speaking, OPP is a point in the manufac- turing value chain for a product where the product is linked to a specific customer order and also divides the manufac- turing stages that are forecast-driven from those that are S. Hemmati :M. Rabbani (*) Department of Industrial Engineering, College of Engineering, University of Tehran, Tehran, Iran e-mail: mrabani@ut.ac.ir Int J Adv Manuf Technol (2010) 48:801–813 DOI 10.1007/s00170-009-2312-4 customer-order-driven [3]. Different manufacturing situa- tions all relate to different positions of the OPP (see Fig. 1). With respect to this concept; the OPP is positioned after the last workstation in the production line in MTS mode. On the contrary, if products are manufactured to order, the OPP is positioned at the beginning of the production line. In hybrid MTS/MTO system, the exact location of OPP is not obvious. There are several factors that affect the position of the OPP in hybrid MTS/MTO. So a compre- hensive methodology is needed to determine the location of OPP in this system considering all driving factors. A schematic of this production mode is shown in Fig. 2. As discussed by Soman et al. [4], since MTS/MTO system is a combination of two production systems, a variety of issues than those required in a pure MTO and a pure MTS system are arising in a hybrid MTO/MTS production situation, and it is difficult to handle all these issues simultaneously. In this regard, a hierarchical decision making is a reasonable approach to solve the issues involved. The pioneer of the application of hierarchical production planning (HPP) in hybrid MTS/ MTO system is due to Soman et al. [4]. The proposed HPP consists of three decision-making levels. These levels are MTS/MTO decision, capacity coordination, and scheduling and control. Although the inputs and outputs of each level are ascertained in this HPP structure, no decision-making models have been proposed to specify the outputs. In this paper, we focus on the first level of the proposed HPP framework by Soman et al. [4] which deals with determination of the appropriate product delivery strategy for different products in a manufacturing system (MTS/MTO partitioning). To make the above strategic decisions in MTS/MTO systems, we propose a decision-making structure considering analytic network process (ANP) methodology. The proposed structure results in three delivery strategy: MTO, MTS, and hybrid MTS/MTO. 1.1 MTS/MTO literature Due to the long use of MTS policy by many firms to meet the demands, the bulk of research regarding product delivery strategy is about how to plan and schedule the products in the production system to meet the forecasted demands. There are considerable amount of works in the literature regarding the production planning and scheduling techniques for MTS systems, especially in material require- ments planning-based systems (e.g., see [5]). Of particular interest, the HPP approach is one of the most applied methodologies for MTS companies because of its several advantages in practice (e.g., see [6]). In contrary to MTS systems, the production system in MTO firms activates only when a new order enters the system. Hence, the production planning and scheduling issues are different from those of MTS firms. The main objective in MTO environments is to manage the delivery dates of arriving orders in order to reach short and reliable delivery dates. To achieve this goal, firms should apply appropriate produc- tion planning [7]. The literature review on MTS/MTO systems reveals that there is only a handful of research regarding production planning and control in these systems. This matter that research regarding MTS/MTO systems is still in its infant stages was taken to consideration by Soman et al. [4] as well. To the best of our knowledge, the thorough work in this regard was carried out by Soman et al. [4]. They proposed a comprehensive hierarchical production plan- ning framework that covers the important production management decisions for MTO/MTS situations in food processing. This framework consists of a three-level decision-making structure. At the first level, the decisions relating to determining which products to manufacture to order and which products to manufacture to stock are taken. Second level is allocation of production orders for both MTO and MTS products to planning periods. At the third level, there are scheduling and control decisions in which the production orders are sequenced and scheduled. Chang et al. [8] developed a heuristic production activity control model to schedule and control wafer manufactur- ing in a hybrid wafer production environment (MTO and MTS). For MTO orders, they developed a rigid order release plan and dispatching control and proposed a method of releasing the orders so as to fill up to an appropriate level For MTS orders. Yingdong [1] devel- oped a mathematical model as a decision tool to design hybrid MTS/MTO systems and searched for the econom- ical base stock level and location necessary to meet specified service constraint. Also, he showed how to determine the optimal point separating the MTS and MTO operations for both balanced and unbalanced flow lines. Rajagopalan [9] proposed a nonlinear integer program with Product delivery strategy Design Fabrication Final assembly Shipment Make To Stock Make To Order OPP OPP Fig. 1 Different product deliv- ery strategies relate to different order penetrating points [3] 802 Int J Adv Manuf Technol (2010) 48:801–813 service level constraints for MTS/MTO partitioning problem. He developed a heuristic procedure to solve this problem. As a recent work, Zaerpour et al. [10] presented a novel hybrid methodology in MTS/MTO manufacturing systems for partitioning the MTS/MTO products. They proposed Fuzzy AHP-SWOT approach as a strategic decision-making meth- odology for partitioning the products. 1.2 ANP literature Decision-making problem, such as making decision on production policy, is the process of defining the decision goals, gathering relevant information, and selecting the optimal alternative [18, 19]. In almost all such problems, there may be so many criteria, and the process of evaluating the efficiency of alternatives is a complicated problem. That is, for many such problems, the decision makers put multiple criteria decision-making (MCDM) techniques into account. In MCDM problems, the primary concern for the decision aid is as follows [11]: – Choosing the most preferred alternative to the decision maker – Ranking alternatives in order of importance for selection problems – Screening alternatives for the final decision There are many MCDM methods which have been developed, such as the elimination and choice translating reality (ELECTRE), the technique for order preference by similarity to ideal solution (TOPSIS), and the analytic hierarchy process (AHP), but these methods do not deal with the interdependence among elements. ANP is a relatively new MCDM method recently introduced by Saaty [12] which can deal with all kinds of dependences systematically. It can be used as an analysis tool in those problems where there are interactions and dependencies among the elements of a system [13]. Since the ANP has these advantages, in this paper, we develop an effective method based on the ANP to help companies to select product delivery strategy. A review of the related literature reveals that ANP has been widely applied in decision-making problems. For example, Yuksel and Dagdeviren [14] presented a process for quantitative SWOT analysis which considers dependen- cy among strategic factors using ANP. Jharkharia and Shankar [15] used ANP for selecting a logistic service provider and Wu and Lee [16] for selecting suitable knowledge management strategy. Karsak et al. [17] consid- ered the interrelationship among customer needs and product technical requirements (PTRs) while determining the importance levels of PTRs in the house of quality in a quality function deployment model. To fulfill this require- ment, they employed ANP. Agarwal et al. [18, 19] used ANP method for combining the criteria, enablers, and dimensions governing the supply chain performance, Genser et al. [18, 19] for supplier selection, and Bayazit and Karpak [19] for identifying the impact of different factors on total quality management implementation. Motivated by the literature discussed above, this paper presents a decision-making structure to decide whether an item should be produced under MTO, MTS, and hybrid MTS/MTO system. To address this decision, the analytic network process is used as a suitable multi- criteria decision-making tool which considers both qualitative and quantitative criteria affecting the prob- lem. The proposed structure attempts to determine the commonalities between products in order to form the product families. Formation of product family and application of the same delivery strategy for each family leads to the facilitation and improvement of the produc- tion planning and scheduling for the existing products, which then specifies a common delivery strategy for the products belonging to the same family by applying ANP method. W.C 1 W.C k1-1 W.C k1 W.C k1 +1 W.C k1 +k2-1 W.C k1 +k2 WorkStation WIP Buffer Material Flow OPP MTO Stage MTS Stage Production Trigger Order Pool Fig. 2 A schematic of production line in a hybrid MTS/MTO system [1] Int J Adv Manuf Technol (2010) 48:801–813 803 The rest of the paper is organized as follows. In Section 2, the proposed decision-making structure applying for MTS/MTO partitioning is illustrated in details. This section consists of a description of different steps of the structure, especially the proposed ANP model. To show the applicability of the proposed structure, a real industrial case study is presented in Section 3. Finally, the concluding remarks are provided in Section 4. 2 The proposed decision-making structure The main business objective of each company is to achieve long-term profitability. To do so, the firms have to consistently produce high quality and specific prod- ucts with competitive unit costs and high service levels (i.e., short delivery time). So they are moving from MTS mode to MTO one, and consequently, they are willing to utilize hybrid MTS/MTO production policy. The main strategic decisions in MTS/MTO systems are MTS/MTO partitioning, i.e., which product or product family can be manufactured to stock, which can be manufactured to order, and which one should be made under hybrid MTS/MTO production policy. To address the above decision, we propose a new decision-making structure including six major steps and apply ANP as a suitable multi-criteria decision-making tool which con- siders both qualitative and quantitative criteria affecting this problem. This method can be used as an effective tool in those cases where the interactions among the elements of a system form a network structure and also can transform qualitative judgments into quantitative values. The proposed decision-making structure is depicted in Fig. 3. This structure consists of the following steps: Step 1: Identifying all products that the firm can manufacture. At first, all products of a manufacturing firm including those produced in previous planning periods and also brand-new products are identified. So a list of existing products in the firm is provided in this step. Step 2: Determining the commonalities among products. In order to form product families, we must identify some grouping criteria that are the key attributes of a product. As mentioned earlier, formation of product family and application of the same delivery strategy for each family leads to the facilitation and improvement of the produc- tion planning and scheduling for the existing products because the production issues are the same for all products belonging to the same family. In this regard, we consider the attributes presented by Galan et al. [20] including all characteristics of products required for grouping them and making product families. Following are these attributes: & Modularity: Modularity can be defined as the degree to which a product is composed of independent modules, without interactions between them, and it is necessary for customized and complex products. The total time of machining operations and manufacturing cost are less in the modular design because of the simplicity of those parts. & Commonality: This can be defined as the measure of how well the product uses standardized part with the view of reducing the total number of different parts. Commonality ensures that a component of a product is shared by two or more products of the same family. & Compatibility: The products in the same family must be compatible. It means that they require the same technical operations, target the same market, and are similar in processing time and manufacturing costs. & Product reusability: This attribute measures the use of existing design configurations while reconfiguring manufacturing elements for a new product type. This feature would be maximized when similar products form a family. & Product demand: In order to have the highest possible utilization rate, it is useful to group products with similar demand trends to select a machining system with similar capacity. The combination of the above attributes reveals the extent of commonality among the existing products. Step 3: Product families’ formation. After determining key attributes for each product, a methodology to group products into families is required. The main purpose of grouping products is to reduce the time needed to manufacture products at the required quality, production volume, and delivery time. There are various methods for the formation of product families such as descriptive procedures [21], mathematical programming [22, 23], cluster-based procedures [24, 25], and artificial intelligence [26]. Among these procedures, we employ the methodology proposed by Galan et al. [20] which is developed for making the product families in reconfigurable manufactur- ing systems. Here, to make a new product, it may be required to reconfigure the manufacturing system. So we prefer to utilize this procedure. At first, for each product attribute including modularity, commonality, compatibility, reusability, and product de- mand, a matrix summarizing the similarity between pairs of products is constructed; so, five matrices are achieved. To obtain a unique matrix comprising the values of interaction among products, paired comparison method (AHP) is used. After implementation of AHP, a weight is obtained for each attributes. Then, the elements of unique matrix are calculated as the sum of the coefficients of each product 804 Int J Adv Manuf Technol (2010) 48:801–813 attribute matrix, which are also multiplied by their corresponding weights. Then, the average linkage cluster- ing algorithm is applied in order to group products together. Products with higher coefficient of similarity are grouped to the same family. Finally, a sub-matrix considering the products grouped as a family is created. This procedure is repeated until all products are grouped into a family. For more details, the interested reader is referred to that of Galan et al. [20]. A thorough review of the aforementioned attributes used to form the product family, it concludes that the products belonging to the same family have the same market, process, and product characteristics, and thus, the same delivery strategy can be performed for all of these products. Therefore, our analysis for MTS/MTO partition- ing ends up at the level of product family and is not carried out for each single product. Step 4: Choosing the product delivery strategy for each product family with respect to the driving factors. One of the difficulties of managerial decisions in manufacturing systems is to choose the right production strategy so that the firms can provide high quality and specific products with competitive unit costs and high service levels. In an MTS system, finished or semi-finished products are manufactured based on anticipating future demands. Since the demand data are often volatile and forecasts are inaccurate, this usually yields unnecessary inventories and also unexpected stock holding costs. With respect to the OPP concept, the OPP is positioned after the last workstation in the production line in MTS system. In an MTO system, an order is fulfilled only when it enters the system. While this system eliminates finished goods inventories and reduces a firm’s exposure to financial risk, it usually results in long customer lead times and large order backlogs [3].