Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1598-7248 (Print)
ISSN : 2234-6473 (Online)
Industrial Engineering & Management Systems Vol.19 No.1 pp.143-163

A System Dynamics Approach Towards Analysis of Hybrid Make-to-Stock/Make-to-Order Production Systems

Moeen Sammak Jalali, S. M. T. Fatemi Ghomi*, Masoud Rabbani
Department of Industrial Engineering & Management Systems, Amirkabir University of Technology- Tehran Polytechnic, Tehran, Iran
School of Industrial Engineering, College of Engineering, University of Tehran, Tehran, Iran
*Corresponding Author, E-mail:
January 27, 2018 July 26, 2018 November 29, 2018


Hybrid Make-To-Stock (MTS)/Make-To-Order (MTO) production system benefits from both pure MTS and MTO systems and thus, is regarded as a valuable production system in order to respond today’s market conditions such as demand uncertainty, shortage of raw materials, and high inventory holding costs. Our purpose is to explore such systems over covering most of the influential factors that have not been taken into account so far. Hence, a system dynamics (SD) model is proposed in this study considering three different series of workstations (MTS, MTO, and MTS/MTO) in a manufacturing firm with a continuous production line. The performance of the developed model is assessed in terms of holding costs as well as product’s delivery lead-time. Moreover, this study considers the impacts of some significant, exogenous variables such as different outlays including operating expenses, holding costs, and the company’s net profit. The results show the superiority of our proposed model in contrast with previous studies in a cost-oriented point of view.



    1.1 Motivation and Significance

    Literature survey of this paper shows that there is a need to dedicate research works to the development of models as well as mechanisms for rendering a logical coordination of capacity to different production environments such as MTS, MTO, and hybrid MTS/MTO. Experts in companies as well as manufacturing arrangements suffer from lack of a systematic structure to help them try different decision approaches and assuring the companies’ board of directors that a policy or set of policies is more practical than the other available ones and will rebound to a better manufacturing and sales situation. With this regard, experts need to be informed of possible negative aspects of their planned strategies. This fact in real industrial environments motivates this study to develop a novel approach for the use of specialists in any manufacturing environments for helping them to select the most appropriate policy with least possible disad-vantages. With this regard, a SD model is presented here including several variables and parameters influencing the company/arrangement in different aspects. The interdependency of variables in a manufacturing arrangement, demonstrates the practicality of using systematic thinking and developing a systematic model as well as surveying its derivation through time by means of Vensim simulation environment.

    1.2 Hybrid MTS/MTO

    As the time goes on, production planning has been turned out to be increasingly significant because of the dynamism, complexity, and the globalization of the economy. Production planning is known as an efficient tool, which helps organizations and corporations in order to respond flexibly to different conditions of markets. In nowadays’ business environments, a manufacturing firm which has the ability to fulfill customer orders quickly, as well as to offer customized products, definitely has the competitive advantage to make profit. Nevertheless, the necessity to have a high product diversity and swift response times, places contradictory demands on the production system (Lee, 1996;Lee and Tang, 1997;Fisher et al., 1999). Having these characteristrics taken into account, selecting an appropriate production system/strategy is by far the most important activity among other essential decisions in a manufacturing establishment.

    Production strategies are classified based on their ability to either decrease the customer lead-time (known as responsiveness) or deliver a more customized products (customization). Hence, from this point of view, manufacturing strategies change from pure MTS with the highest level of responsiveness, to pure MTO including uppermost level of customization (Meredith and Aknic, 2007). Figure 1 explicitly expresses that between these two strategies, there are various production approaches to meet both customization and responsiveness with an appropriate proportion to each, based on the organizations’ goal and manufacturing processes. The main discrepancy between MTS and MTO is the timing of the receipt of the customer order relative to the final assembly of the finished product. In an MTS environment, the product is assembled in anticipation of future orders and stored in the finished goods inventory (Youssef et al., 2004), whereas in an MTO system the customer order is received before assembly of the final product.

    A noteworthy proportion of the researches prior to 1990 in the production planning area were aimed at the needs of MTS companies (Hendry and Kingsman, 1989). Nowadays, the choice between MTS or MTO system for a manufacturing corporation is a strategic one. Firms are trying to analyze different working situations in order to make the best choice and to be more competitive in the ever more intense global economy. The main advantage of MTS systems is the shorter lead-time, since the final products are already in stock even before receiving the customer order. In an MTO system, the lead-time may include design, procurement, manufacturing, final assembly and shipment stages, while for a strictly MTS system the lead-time only consists of the shipment period. In a pure MTS environment, the firm’s logistics management plays a significant role in maintaining the competitiveness of the company by determining the variety, size and location of the finished inventories. The main challenge for a MTS system is that it must compute an accurate forecast of customer demand before production planning; otherwise, the market mediation costs will be so high if a overstocking or stock-out happens. Moreover, in a pure MTS environment, products cannot be customized because demand is met from finished stock. In highly competitive industries like electronics and microcomputers, this inflexibility makes MTS less attractive for the manufacturing companies.

    Hence, the inclination toward increased product variety and sweeping changes in market demand have necessitated manufacturing systems to attain the ability of producing both MTO and MTS products (Tsubone et al., 2002). Whereas MTS and MTO represent the two ends of a spectrum, the hybrid MTS/MTO systems are even more useful in practice. Some semi-finished products can be maintained at one or more stocking points in the hybrid systems towards reducing the response time, since the manufacturing delay is just the time needed for the MTO stage if we start with the semi-finished products. A typical MTS system releases raw materials based on a forecast of the future demand. Since forecasts are often inaccurate, either this will rebound to unnecessary inventories or unexpected stock outs. The key benefit of an MTO system is that the release of production only happens after order arrivals, which can prevent inventory explosion. A proper combination of MTO and MTS could have exploited the compensations of both lower inventory and shorter lead-time.

    Another advantage associated with a hybrid MTS/ MTO system is the opportunity to exploit the benefits of delayed differentiation (DD). Delayed differentiation means, “delaying the point in time at which a product assumes its unique identity” (Lee and Tang, 1997). Usually DD systems consist of two stages: a common stage and a differentiating one. At the common stage, generic merchandises are produced and stored close to their final markets. Following receipt of order at the differentiating stage, the generic products are post-manufactured and customized to meet the clients’ demands. Furthermore, DD systems can be used to satisfy the specific prerequisites of various customers. DD strategy also permits specifying only the less variable aggregate orders at the beginning of the replenishment process in order to commit resources toward individual products at the end of the first stage. Since the differentiation decision is made close to demand realization while market information is more accurate, DD provides a hedge against the market uncertainty. The later benefit is also known as learning effect. Recent literature shows several examples in which DD strategy has been successfully implemented to control inventory levels while managing high service levels (Lee, 1996;Swaminathan and Tayur, 1998). Therefore, there are different types of hybrid MTS-MTO systems applicable in practice. They differ from one another in terms of layout, material flow, and inventory control policy used. MTS systems have been traditionally viewed as entirely distinct from and incom-patible with MTO philosophy. Some researchers have addressed the issue of combined MTO and MTS systems, in which some of the products (e.g. type B) are produced in MTO patterns while others (type A) are produced in MTS environment.

    The greatest constant of modern times is change. Accelerating modifications in technology, popu-lation, and economic accomplishments are transforming the world, “from the prosaic to the profound” (Sterman, 2000). System dynamics is a perspective and set of conceptual tools enabling us to understand the structure and dynamics of complex systems, as a rigorous modelling method that facilitates building formal computer simulation of multifaceted systems and using them in order to design policies that are more effective and productive for organizations. In other words, SD is an operative method to enhance learning in complex systems. Learning about complex dynamic systems requires something more than technical tools to create mathematical models. SD is fundamentally interdisciplinary, grounded in the theory of nonlinear dynamics and feedback control developed in mathematics, physics, and engineering. Forrester (1961) who developed system dynamics in 1950s pointed out “all decisions are based on models, fre-quently mental models.” Therefore, surveying the appli-cation of SD models in manufacturing environments sounds to be advisable.

    While the production environments are becom-ing more and more complicated, manufacturing strategies have been in progress over the years. The use of various technologies has forced companies and organizations to use a suitable strategy, which has responsiveness as well as a good level of customization at the same time. Consequently, the use of hybrid MTS/MTO has become more and more ordinary during the 20th century. In this regard, a SD model is proposed through this study in order to evaluate the influences of various factors on a hybrid production system. The remainder of the paper is as follows: section 2 reviews literature body of MTS/MTO production systems and production planning through SD methodology. Then, the problem is described through dynamic modelling presented in section 3, whilst section 4 comprises our formulation methodology and base run simulation results. The performance of the proposed model is discussed in section 5 using sensitivity analysis. Finally, conclusions and some directions for further studies are given in section 6.


    The idea in this section is to survey some select-ed articles in order to provide an overview of the extant body of literature. Different approaches have been adopted so far to study different aspects of hybrid MTS/MTO production systems under diverse circumstances. Hence, we first look at what has been done in the field of hybrid MTS/MTO environment using different approaches up to now. Moreover, we investigate some recent studies about the use of SD approach in the field of production planning.

    2.1 Hybrid MTS/MTO Literature Body

    A majority of the operations management re-searches characterized production systems as either MTO or MTS. Hence, there are only a handful of research articles explicitly addressing the combined MTS/MTO situation in spite of being interesting for experts and academicians. One of the first studies towards hybrid MTS/MTO is due to Williams (Williams, 1984) which dealt with many questions raised by MTO products. He represented a single-stage system with stochastic demands and interactions between demands and capacity in order to answer several questions associated with hybrid environments. Rajagopalan (2002) characterized a heuristic procedure to solve a non-linear, integer programming formulation of the problem which specified the MTS/MTO partition and the batch sizes for the MTS items. Bemelmans (1986) considered decomposition of items into slow and fast movers and deliberated the situation as a capacitated single-machine, multi-product for production planning problem. The performance criterion in his research was to minimize sum of inventory holding costs and stock-out costs. Nevertheless, the idea could be further extended by considering extra inventories for MTS in order to have more available capacity for MTO. On the other hand, Li (1992) investigated the impact of customer behaviour and market on MTO/MTS partitioning. Assuming single product as well as stochastic demand and price, quality, delivery, and also lead-time variations, he supposed a condition in which the firm may not ge all the orders and maximized profit using stochastic optimization with infinite time horizons as performance criterion. Nguyen (1998) addressed a combined MTS/MTO situation modelled as a mixed queuing network and used the heavy traffic limit hypothesis in developing the process of finding estimates of fill rates and average inventory levels.

    Supposing to have no set-up times and costs, Carr et al. (1993) represented exact expressions for cost of a strategy mentioned as an example of MTO/MTS situation. Regarding unit demand with stochastic arrival time, they minimized sum of invrntory holding costs and stock-out costs by means of using a M/D/1 queue with two priority class; preemptive resume between priority class and First Come First Served (FCFS) within a class. Adan and Van der Wal (1998) investigated the effects of combining MTO and MTS on the production lead-time in single and multi-stage manufacturing environments. They assumed two types of products with stochastic demands, exponential processing times, and exponential times between coming orders. Gupta and Benjaafar (2004) proposed models to compute the costs and benefits of DD in series production systems, where the order lead times are load dependent. They supposed that switching from assembling one product to another does not need any further setups. Customer orders in the represented models were expected to meet on a FCFS basis. While Gupta and Benjaafar (2004) considered unit processing times at each stage exponentially distributed, Chang et al. (2003) developed a heuristic production activity control (PAC) model to achieve different production criterion (for MTO and MTS) in a hybrid production environment. They assumed that MTO and MTS orders are distributed uniformly during the planning period. In order to simplify the problem, they also supposed that only one kind of product of MTO orders exists. Van Donk (2001) has associated the thought of manufacturing strategies with the order decoupling point (OPP) through presenting some criterion such as set-ups, capacity and so on, by means of having planning and scheduling purposes taken into consideration. For example, high set-ups and an orientation to use capacity as much as possible, results in planning long production runs and stocks of finished products. Olhager (2003) correspondingly offered an extension to that of van Donk (2001) by defining more measures and criteria. Alp and Tan (2006) considered a MTS production environment with flexible capacity and stochastic demand. On the other hand, Ghalehkhondabi et al. (2017) developed an inegrated decision making model for pricing and locating customer order decoupling point in a MTS/MTO supply chain.

    As mentioned before, there are different approaches which can be applied to the field of hybrid MTS/MTO production systems. One of which, is hierarchical pro¬duction planning (HPP) that had first been familiarized by Hax and Meal in the 1970s. With this methodology, a decision could be made by means of diverse decision making levels with dissimilar characteristics (Hax and Meal, 1975). The application of HPP in such environments was first surveyed by Soman et al. (2004) who defined three decision levels including strategic, tactical, and operational ones. In the first level which is so called as MTS/MTO decision, product families are formed; then the manufacturing systems and OPP-related locations are decided. The second level, known as capacity coordination, assigns capacity, MTS, MTO, and hybrid orders. Here orders are either accepted or rejected on the basis of their profitability. Consequently, cost-effective MTO and hybrid orders are accepted and the non-profitable ones are rejected in order not to stock up capacity. The last level (scheduling and controlling) consists of determining the sequence and in-depth production plan of the shop to meet due dates and lot sizes obtained in the previous level. Soman et al. (2007) tested the conceptual production planning and inventory control framework for combined MTO and MTS. Rafiei and Rabbani (2012) addressed capacity coordination for hybrid manufacturing systems including three types of products: pure MTS, pure MTO, and hybrid MTS/MTO. This level of HPP includes some substancial factors such as order acceptance/rejection pro¬gramme, order due-date setting, lot-sizing of MTS products, and also determining prerequisite capacity during the planning horizon of the level. They represented a five-step model attempting to decide on several significant decisions in hybrid manufacturing systems in the tactical level of HPP. With this regard, they presumed the inventory of raw materials to be enough for meeting MTO orders in spite of pointing out some limitations in the proposed model. For example, the considered model was settled for production systems without any buffer and the shop-floor layout assumed to be as job shop. Zaerpour et al. (2008) presented a strategic decision-making structure by a novel hybrid methodology to determine whether a specific product should be produced with MTO or MTS strategy. Rafiei et al. (2013) addressed the second and third levels of a HPP approach to hybrid MTS/MTO production systems, proposing a structure consists of both mid-term and short-term production planning levels.

    Renna (2016) developed a simulation based metho¬dology to test the production control strategies in order to release MTO and MTS orders. He explained that the proposed approach leads to better results for services level and reduces the MTS level in various conditions. Nagib et al. (2016) reveiwed previously published papers in the field of inventory control model in the food and beverage processing industry. They discussed four inventory models including MTS, MTO, economic order quantity (EOQ), and hybrid MTS-MTO models and highlighted the important role of the explained models in food processing industries. Beemsterboer et al. (2016) developed a Markov Decision Pocess model for a two-product hybid system with the intention of determining when to manufacture MTS and MTO products. They characterized optimal policies and explained the factors influencing the decision. Ghalehkhondabi et al. (2016) considered demand uncertainties in two echelons of a hybrid MTS/MTO manufacturing supply chain. They obtained the number of processes bases upon MTS, the order quantity of the decoupling point, and the order quantity of final customers. Beemsterboer et al. (2017a) proposed four methods of integrating MTS items into job shop control and evaluated them by means of discrete event simulation. They explained that their method of “smart integration” reduces MTS lost sales in comparison with always prioritizing MTO. On the other hand, Beemsterboer et al. (2017b) used Markov Decision Process modelling and showed that optimal policy for Hybrid MTO/MTS systems varies the lot size. They evaluated the performance of flexible lot sizing policy for a two product hybrid system. Rafiei et al. (2017) addressed a single- station production system with two parallel machines and developed a queuing model. They conducted numerical analyses with respect to the exclusive performance measures for MTS and MTO products.

    Rabbani and Dolatkhah (2017) simulated the pro¬cesses of a five-star restaurant using a dicrete event simu¬lation based method. They investigated the simultaneous production planning of MTS and MTO products and optimized the important parameters fom the perspective of restaurant management. Halawa et al. (2017) presented a case study focusing on the transition to a hybrid MTS/MTO in an automotive manufacturer. They used Discrete event simulation to obtain the results of the project and showed that applying hybrid strategy reinforces the responsiveness of the supplyy chain to unpredictable customer demand. Jia et al. (2017) proposed a production model for flexible flow shops with the intetnion of reducing the total inventories, earliness, and tardiness. Their proposed model divides a flexible flow shop into a MTS part and a MTO part b means of applying a decoupling point. Arima et al. (2017) discussed real time solutions to product- mix scheduling poblems in MTS and MTO production systems. The objective was to maximize resource utilization as well as keeping a due date and Q- time restrictions of every production lot.

    2.2 The application of System Dynamics in Production/Inventory Environments

    Although there are vast spectrums of articles written in the field of developing dynamic models for production/inventory problems, the researched papers published in business scopes are quite limited. Neverthe-less, we investigate some of the related articles; Lee and Chung (2012) analyzed inventory policies and the dynamic relations between inventory costs by means of developing a SD model. They assumed that the demands could be changed constantly. Some other important assumptions of the mentioned article are: known order cycle, considering single item, lot-size is the same for each delivery, shortages forbidden, deterministic demand and production rate. Then a mathematical model was represented and finally, a SD model had been developed. The results indicated that the SD model and simulation are useful instruments, which can provide a more thorough, robust, and long-term perspective solution. The simula-tion results verified the propriety of SD simulation meth-odology in order to apply at time evolution related re-search subjects.

    Georgiadis and Michaloudi (2012) integrated the advantages of operations research, control theory, and simulation disciplines into a real-time production planing and control system for job-shop manufacturing in order to develop a comprehensive dynamic model using the SD methodological approach, which is proved to be appro¬priate for studying the dynamic behaviour of complex manufacturing systems. With this strategy used in practice, they have studied the system’s responsiveness and proposd a SD model. Poles (2013) modelled a production and inventory system for remanufacturing items by means of a SD simulation modelling approach in order to discover the dynamics of a remanufacturing process and evaluate system improvements strategies. The remanu¬facturing procedure has been demonstrated considering some key factors such as integrated remanufacturing/ production capacity, lead times, backorder, and inventory coverage. Developing the dynamic model, the author propounded a case study in order to assess the findings attained from the simulation analysis and to further validate the proposed model. Finally, the author utilized the total cost of production, inventory, and backorder as measures of performance (Anissa, 2017), which were employed in order to evaluate the simulation results. For example, one of the results derived from the article was that an increase of the production lead-time rebounds to higher effects on the system’s performance in comparison with an equivalent increase in the manufacturing lead-time.

    Searching approximately, most of the scientific databases have not resulted in successfully finding enough related studies in combining SD approach with the hybrid MTS/MTO production environments. In fact, we may conclude that there are very few research papers in relation with the analysis of hybrid MTS/MTO manufacturing systems using SD methodology. The only available article is due to Rafiei et al. (2014), which dealt with effects’ evaluation of different capacity coordination mechanisms on the system’s performance, by means of representing a SD model. The main assumptions the authors pointed out were encompassing three demand types (MTO, MTS, and MTS/MTO products), continuous production environment, segregated demands based on their analogous OPPs, constant production capacity throughout the planning horizon, multi-stage production system, and limited capacity of workstations. Finally, the performance of the proposed system was assessed in terms of system’s delivery lead-time. Despite having several features of hybrid production environments taken into consideration, they have not evaluated some important characters of capacity coordination in hybrid systems. For example, pricing and profit maximization, which are nowadays considered as most important factors influencing the companies’ judgments in any circumstances, have not been considered and evaluated in their proposed SD model. Hence, in this study we will focus on covering such impacts in our proposed SD model. Although the previous study considered only a hybrid production system with three workstations, we consider three different workstations including a set of MTS, MTO, and MTS/MTO workstations. Inventory holding costs have not been studied in the mentioned paper, while we consider the impacts of various inventory circumstances. There was no differences between the conditions in which either customer or the firm’s order acception or rejection takes place, whereas we consider the potential impacts of various factors in our developed model. Although processing times have not been taken into account in MTO workstations, we modelled the mpacts of various processing times exogenously.


    In this section, the problem is defined by rendering a dynamic hypothesis based on closed-cycle-systems. Then, we nominate some factors that have an influence on the model. Henceforth, the central loop of the proposed model is illustrated. Then, some branches from the central loop are demonstrated and finally, we will represent the whole model with the corresponding enlightenments. We should mention that the process of choosing factors in the presented system dynamics model is based on several aspects. Some of which have been explained throughout the third section. Some others are explained throughout the review section. In other words, the logic of choosing factors for the developed model is on the basis of the structure of the surveyed systems (here, the structure of hybrid MTS/MTO systems) and the industrial and research documents in the literature. We have added this point in the literature and highlighted it. As mentioned before, there are a set of three different workstations in a company, which together affects performance of the establishment For instance, an increase in the overall capacity of the company (affected by the three categories of workstations) will provide the company accept more customer orders. The company’s order acceptance is a somewhat influential factor that will rebound to an increase in the net profit. On the other hand, the company will be convinced to increase the overall production, which will eventually leads to capacity increase. The central loop is displayed in Figure 2.

    Each of the four factors mentioned in the central loop are affected by some other causes. For example, the overall production capacity is affected by overall capacity before Order Penetration Point (OPP) and after OPP. The utilization of capacity is profoundly dependent on the received customer orders and it is shown in practice that orders are not arriving continuously (Samadhi and Hoang, 1995). On the other hand, the customer suggests a specific time for the products delivery, which is known as desired delivery lead-time. Another effective concept to the mentioned factor is the actual delivery lead-time. If the ratio between actual and desired delivery lead-time (which is called “AD ratio”) is bigger than one, orders will be rejected for MTO and MTS/MTO products or customers will give up their purchase in MTS products. Hence, the company would accept incoming orders based on the available orders and the AD ratio.

    On the other hand, delivery and manufacturing lead-times depend on the total backlog (TB) and planned backlog (PB) (Meredith and Aknic, 2007), in which the TB is influenced by planned and unplanned backlogs. From this point of view, there are two factors affecting the net profit: Label Price and Penalty (although there are more influential factors that will be explained in the final model).

    On the other aspect, customers giving orders typically quote a due date and a deadline and may penalize the manufacturer for late deliveries. This, rebounds to reduced revenue and even loss customers for the manufacturer (Oğuz et al., 2010). Therefore, the penalty is defined as the product of AD ratio and unit delay penalty. The overall production will also affect the available production capacity, which will rebound to a decrease in the actual delivery lead-time, entailing the company accept more customer orders. The available production capacity will increase the shipment rate, proliferating demand Responses towards the total demand. The total demand effects the customer order and then, rebounds to unplanned backlog growth, affecting the total backlog. The dynamic model of the mentioned descriptions has been depicted in Figure 3.

    As mentioned in the previous section, three types of workstations are taken into account in the pro-pounded model. The total expected demand is divided into these three types, which are dependent on specific proportions for each. Hence, the production lot-sizes for each type will be specified. Note that these quantities are before the OPPs. Then, the available production capacity is affected by before OPP production items.

    Thereafter, the production quantities after OPP are calculated based on the available capacity and the total demand (considered as overall demand of all three types that the customers states after OPP). The explained cause and effect relationships will finally affect the central loop and its elements such as overall production and overall capacity. It might appear, at the first glance, that when demand is greater than the capacity, the firm should expand capacity in order to be able to increase its response rate (Sridharan, 1998). There are various characteristics for MTS, MTO, and MTS/MTO production systems. As we combined these three types into segregate workstations and placed it in a single company, their different aspects sometimes oppositely affect the parameters of manufacturing systems. For example, when the company has more MTO orders than MTS ones, the holding costs will be effectively lessened (since we do not need inventories for MTO products). If the firm is short of capacity in one period, then excess demand is backlogged and carried over to the next period.

    However, if the capacity is well-balanced re-garding to the demand, then there is significant opportunity in order to decrease the expected backlogs by revising the allocation periodically. MTO production using flexible capacity is often associated with the ability to make a wide variety of products on-demand. Such capabilities are needed when, for instance, customized products are offered. However, when market demand is primarily for standard products, a firm might traditionally view this as a situation calling for efficient capacity and MTS production. Nevertheless, whether an item is MTO or MTS depends not only on the items’ demand but also on their unit processing times, holding costs, and setup times. Items with higher per unit holding costs, are more likely to be considered as MTO for this rebounds to a reduction in total holding costs. On the other hand, items with higher setup times are more likely to be MTS. Since MTO pro-duction of such items, requires numerous setups and uses up lots of capacity (Rajagopalan, 2002).

    Here we define three capacity related costs (Allon and Zeevi, 2008):

    • K = Cost of Adding a Unit of Capacity

    • k =  Return from Selling a Unit of Capacity

    • h =  Holding Cost Per Unit

    If the ratio between capacity cost difference (K-k) and the holding cost (h) are high, we expect the firm to restrict use to inventory in order to meet variability in demand. Considering MTS, MTO, and MTS/MTO required capacities after OPP, their total required capacity affects the company’s overall production, rebounding to an increase/decrease in the available and overall production capacities. Furthermore, after developing the model, it is implied that there are two more factors affecting the net benefit. For example, the company’s decision about whether to accept MTS/MTO orders and how much, has a great influence on the net profit. Another influential factor on net benefit is order acceptance of MTO systems. As mentioned before, more MTO products rebound to holding costs decrease. This will have impacts on the operating expenses and, as a result, causes to suggest a lower label price, rebounds to increasing the net profit, overall production, and overall capacity. Finally, the final model has been depicted in Appendix B. In the next sections, the description of the whole model’s formulation, the results and analysis of them are illustrated. The developed Stock and Flow Diagram (SFD) toward the represented model for our proposed hybrid MTS/MTO production environment is settled as displayed in appendix C. However, we focus on elucidating our final causal loop diagram since all the variables in the proposed system are considered through time and the apprehension of causal loop diagrams are better due to their simplicity.


    As mentioned before, three types of demands including MTS, MTO, and MTS/MTO demands are studies in this research, which are considered respectively as following:

    10+RANDOM UNIFORM(0, 2, 0) × SIN(1 × (Time))

    RANDOM NORMAL(3, 9, 6, 2, 0) +2 × SIN(0.3 × (Time))

    RANDOM NORMAL(2.5, 9, 5, 3, 0 ) +2.5 × SIN(2 × (Time))

    Actually, the demand functions are a combina-tion of probability functions and the fluctuations over time. With this method of formulating demand types, the model will have the flexibility to respond to variable demands and demand fluctuations. Having the highest average in quantity, Figure 2 indicates that the MTS demand (through time) has the least fluctuation in comparison with the two other ones and therefore, is more predictable. On the other hand, demand fluctuation in MTO and MTS/MTO demands are more than that of MTS. Hence, they are less predictable with lower average quantities in contrast with MTS demand.

    The total demand forecast is calculated through sum of the MTS, MTO, and MTS/MTO demands. Then, the capacity share of each workstation is determined through equations (4) to (6). Figure 3 shows the capacity shares of each demand type workstations.

    Capacity Share of MTS Workstations = I n t e g r a l ( ( M T S  Demand / Total Demand) Capacity Share of MTS Workstations)dt

    Capacity Share of MTO Workstations = I n t e g r a l ( ( M T O  Demand / Total Demand) Capacity Share of MTO Workstations)dt

    Capacity Share of MTS/MTO Workstations = I n t e g r a l ( ( M T S / M T O  Demand / Total Demand) Capacity Share of MTS/MTO Workstations)dt

    On the other hand, MTS production is dependent on holding cost. If inventory holding cost is high, the system should reduce its MTS production and, consequently, higher holding costs lead to less MTS production until the company decides not to produce such products any more. Therefore, as mentioned before, the firm is projected to restrict using inventory under such situations, and the MTS production will lead to zero. This is expressed in the model for calculating MTS production by means of the following formulas:

    KH Ratio = Capacity Cost Difference / Holding Cost Per Unit

    We first need to define the “Capacity Cost Dif-ference” in order to reach to the estimation of “KH Ratio”, which is determined through the following equation:

    Capacity Cost Difference = ABS( Cost of Adding a Unit of Capacity Return from Selling a Unit of Capacity)

    In which “Cost of Adding a Unit of Capacity” is assumed to have normal distribution, using the following parameters:

    Cost of Adding a Unit of Capacity = RANDOM NORMAL(30, 70, 45, 20, 7)

    Moreover, “Return from Selling a Unit of Ca-pacity” is estimated using the following formula:

    Return from Selling a Unit of Capacity = SMOOTH(Net Profit / Available Production Capacity, 2)

    The holding cost is assumed to be constant with 7.5 $ per unit of capacity. Accordingly, the MTS production is estimated through equation 11. The equation exemplifies that when the “KH Ratio” is less than one (because of higher holding costs), the MTS Production will lead to zero and the company terminates responding to MTS demands. Figure 4 shows the outcome of simulation for the MTS production when the holding cost is not so high. It is concluded that some demands are being rejected in the periods where the capacity cost difference was less than the specified holding cost.

    MTS Production = IF THEN ELSE (KH Ratio<1, 0, MTS Production Coefficient × SMOOTH(Capacity Share of MTS Workstations, 6))

    The MTO production is also estimated based on its pre requisite capacity and the processing time, which is displayed in equation (12).

    IF THEN ELSE(MTO Order Acceptance  by Company<MTO Required Capacity,  IF THEN ELSE(Total Processing Time Per Unit <MTO Delivery Lead Time, UB × SMOOTH (Capacity Share of MTO Workstations, 3), 0), LB × SMOOTH(Capacity Share of MTO  Workstations, 3))

    To model the MTO production, the structure shown in Figure 5 is developed. The required capacities for MTS, MTO, and MTS/MTO demands are determined through the following equations:

    MTS Required Capacity = I n t e g r a l ( MTS Expected Demand MTS Required Capacity)dt

    MTO Required Capacity = I n t e g r a l ( MTO Expected Demand MTO Required Capacity)dt

    MTS/MTO Required Capacity = I n t e g r a l ( MTS/MTO Expected Demand MTS/MTO Required Capacity)dt

    In which the expected demands are calculated using SMOOTH function, e.g. for the one related to MTS, it is indicated as following:

    MTS Expected Demand = SMOOTH(MTS Demand, 10)

    Furthermore, total processing time is expected to be constant during the planning horizon of the study. The MTS/MTO Production is also structured through the equation (17). Figure 6 expresses the behavior of MTO and MTS/MTO production in the developed model:

    IF THEN ELSE ("MTS/MTO Order Acceptance by Company" <"MTS/MTO Required Capacity",  UB × SMOOTH("Capacity Share of MTS /MTO Workstations", 10), LB × SMOOTH("Capacity Share of MTS /MTO Workstations", 10))

    Then, before and after OPP productions are premeditated respectively through workstations’ production and total required capacity, using exponential delay according to equations (18) and (19). Figure 7 demon-strates the simulation result for these two variables. The required capacities for each type of production in the base run are also depicted in Figure 8.

    Before OPP Production = SMOOTHI(Workstations Production, 4, 5)

    After OPP Production = MAX(SMOOTHI( Total Required Capacity Available Production Capacity, 4, 2), 0)

    Considering products’ price, penalties for deliv-ery lateness and calculating net profit are some of the most significant modules in our developed model, which has not been studied so far in the previous studies. In the base run, a label price has been defined for the products through the following equation:

    Label Price = (1+Expected Margin) × Operating Expenses

    The expected margin is supposed to be constant during the planning horizon. However, operating expenses are the sum of holding cost per unit and fixed operating costs, which is normally distributed with the following parameters:

    Fixed Operating Cost = RANDOM UNIFORM(20, 50, 6)

    Alternatively, penalties are determined through multiplication of the differences between actual delivery lead-time, desired delivery lead-time, and the unit delay penalty, which is supposed to be constant (30$). Consequently, equation (22) determines penalty and finally, the net profit is estimated through equation (23), based on the total order acceptations. Figure 9 indicates penalties and label prices for the base run and net profits are demonstrated in Figure 10.

    Penalty = IF THEN ELSE (AD Difference>0, AD Difference × Unit Delay Penalty, 0)

    Net Profit = I n t e g r a l ( (Label Price Penalty) × Company Order Acceptance Net Profit) d t

    Delivery lead times are the other important fac-tors considered in the developed model, which are based on order fulfillment rate, calculated through equation (24):

    Order Fulfillment Rate = SMOOTHI(Total Expected Demand × Shipment Rate, 5, 6)/100

    The total expected demand is calculated based on MTS, MTO, and MTS/MTO expected demands, and the purchasing factor, using the following formula:

    Total Expected Demand = I n t e g r a l (IF THEN ELSE(Purchasing Factor>0,  SMOOTHI(MTS Expected Demand +MTO Expected Demand +"MTS/MTO Expected Demand", 3, 6),  DELAY1(MTS Expected Demand +MTO Expected Demand+"MTS/MTO  Expected Demand", 5)) Total Expected Demand) d t

    In which “Purchasing Factor” is estimated through the subtraction between customer desired price and label price, which is determined in equation (26):

    Purchasing Factor = Customer Desired Price-Label Price

    Delivery lead times are assumed to have random normal/uniform distributions multiplied with the order fulfillment rate. Therefore, the following formulas illustrate delivery lead times:

    MTS Delivery Lead Time = RANDOM UNIFORM(1.3, 4.8, 3) × Order Fulfillment Rate

    MTO Delivery Lead Time = RANDOM NORMAL(4, 6, 5, 1, 2) × Order Fulfillment Rate

    MTS/MTO Delivery Lead Time = RANDOM NORMAL(4, 6, 5, 1, 3) × Order Fulfillment Rate

    In order to calculate the actual delivery lead-time, weighted average of products’ delivery lead-time is adopted. To do so, weights of four, one, and two are applied for MTO, MTS, and MTS/ MTO products, respectively.


    The red variables with brown arrows stand for exogenous variables, which are our primary instruments for sensitivity analysis and assessing system’s response under various conditions. Exogenous variables are those from outside the boundary of a system that unexpectedly influences the performance of the whole system. So, this kind of variable is the best choice for analyzing the per-formance of the developed SD model to ensure the rectitude of model building. In other words, we can mention that a SD model is incomplete without exogenous variables as it is necessary to investigate both endogenous and exogenous ones. With the intention of ensuring the logical behavior of the proposed model, consider “Holding Cost per Unit” as an exogenous variable. Its value has supposed to be 7.5 $ per unit in the base run. Consider holding cost a big number, for example 70$ per unit. As mentioned before, the system should (is expected to) reject MTS demands while the holding cost is high. Figure 11 Figure 12 explicitly support this fact. The company has rejected all MTS demands since it has restricted using inventories because of high holding costs.

    The similar results are obtained by applying Ramp (linear) increase and Step increase in the holding cost, which is displayed in Figure 13 Figure 14. For in-stance, equations (30) and (31) are used to generate holding costs:

    EXP(2)+RAMP(EXP(2), 10, 50)

    EXP(2)+STEP(EXP(4), 10)

    From another viewpoint, the system’s perfor-mance is evaluated through “total processing time per unit”. This factor has been assumed 0.05 day in the base run. Now considering 0.4 as a new value for this factor, the system is expected to reject MTO demands since the more the value of the processing time, the more probable exceeding the delivery lead-time, rebounding to an increase in the rejection rate of MTO product demands. Similar results are achieved while considering ramp and/or stepwise increase for the processing time.

    Finally, the system’s response is assessed to-wards demand uncertainty. Here two types of changes in demand are accomplished, including linear (ramp) increase and step increase. Consider using the following equations to generate MTS demand:

    10+RANDOM UNIFORM(0, 2, 0) × SIN(1 × (Time))+RAMP(0.25, 0, 96)

    10+RANDOM UNIFORM(0, 2, 0) × SIN(1 × (Time))+STEP(5, 40)  

    Applying demand uncertainties, the system’s responses are displayed in terms of delivery lead-times for MTS and MTO products, respectively. As shown in Figure 15.

    Figure 16, MTS delivery lead-time increases step wise and linearly, because the lowest priority has been considered for MTS products, while the delivery lead-time for MTO products has not been changed no-ticeably, because of its highest priority. The priority order assumed in this research is the following:

    Moreover, according to what mentioned before, weights of four, two, and one are applied for the pointed-out products correspondingly. Furthermore, the superiority of the proposed model is shown by means of comparing the present work with the literature. Table 1 compares our proposed model with other recent studies in the literature.


    The application of hybrid MTS/MTO production environments are becoming more and more dominating because of its flexibility against different demand situations. Despite lots of papers working in the field of MTS/MTO, literature survey of this paper shows that there is a need to dedicate research works to development of models and procedures for investigation of capacity coordination dynamics in hybrid manufacturing establishments. Hence, this paper proposed a system dynamics model for a hybrid MTS/MTO production environment with three different series of workstations (MTS, MTO, and MTS/MTO).

    The key contribution and significance of this study is threefold. First, unlike most studies considering only a hybrid workstation that responds to all demand types, this paper defined three different workstations for responding to MTS, MTO, and MTS/MTO demands separately. This feature enables establishments to handle demand uncertainties and fluctuations in customer orders independently. Second, this is the first study that investigates the impacts of pricing and profit maximization in hybrid MTS/MTO production environments. Third, with reference to the functional and contextual analysis of hybrid systems, the dynamism of such systems has been explored considering the most influential factors in the proposed SD model. Finally, the system’s performance was assessed by analyzing reactions of the propounded model under different conditions such as demand uncertainty, variable operating expenses, and pricing. The sensitivity analysis confirms the logical behavior of our developed model and, therefore, verifies its superiority in contrast with the previous study, since it considers factors that are more influential in order to explore dynamism in hybrid manufacturing environments.

    However, any simulation model like our repre-sented SD model is inherently limited from the viewpoint of modelling parameters being investigated in comparison with ignored parameters. Hence, three suggestions can be considered as further studies in this area. First, analyzing the effects of labor development and different preoccupation strategies such as hiring/firing policies are worthwhile in order to study their special effects on the developed SD model. This issue is on our research line. On the other hand, the most competitive aspect among establishments in nowadays’ global trade and economy is quality. In addition to all the limitations mentioned above, we have not established an appropriate stability analysis of the proposed system. Therefore, considering qualitative factors as well as quantitative ones and analysis of their impacts on different demand types, is our second suggestion for future research directions. Third, the proposed SD model can further be improved in different ways. For example, processing times have been considered exogenously in our proposed model, while it can be investigated endogenously. The same procedure can be implemented for additional costs of increasing a unit of capacity and unit shipment time.



    Production strategies to meet customization/responsiveness (Meredith and Aknic, 2007).


    Basic structure of capacity increase in the proposed model.


    Dynamic Structure of OPP.


    Simulation results for MTS production in base run.


    The structure of “MTO Production” in the model.


    Simulation results for MTO and MTS/MTO production in base run.


    Simulation results for before and after OPP production in the developed model.


    Simulation results for MTS, MTO, and MTS/MTO capacities in the base run.


    Simulation results for label price and penalty in the base run.


    Simulation result for net profit in the base run.


    KH Ratio, before and after changing the holding cost.


    MTS Production, before and after changing the holding cost.


    System’s behavior in MTS Production against linear increase in holding cost.


    System’s behavior in MTS Production against step increase in holding cost.


    Simulation results under demand uncertainty for MTS delivery lead-time.


    Simulation results under demand uncertainty for MTO delivery lead-time.


    Sample SFD of the proposed hybrid MTS/MTO production environment.


    Causal loop of the developed model [Final model].


    Comparing our model with other proposed models in the literature

    Description of terms and functions applied in the Vensim modelling environment


    1. Adan, I. J. B. F. and Van der Wal, J. (1998), Combining make to order and make to stock, Operations Research Spektrum, 20(2), 73-81.
    2. Allon, G. and Zeevi, A. (2008), A Note on the Relationship Between Pricing and Capacity Decisions in Make-to-Stock Systems, Columbia University, 1-23.
    3. Alp, O. and Tan, T. (2006), Tactical Capacity Management under Capacity Flexibility in Make-to-Stock Systems, Technische Universiteit Eindhoven, 1-26.
    4. Anissa, R. (2017), Supply chain management sustainability through the performance improvement of mts-mto production system typology, Jurnal Manajemen Industri dan Logistik, 1(1), 58-65.
    5. Arima, S. , Zhang, Y. , Akiyama, Y. , and Ishizaki, Y. (2017), Dynamic scheduling of product-mix production systems of MTS and MTO, 2017 Joint International Symposium on e-Manufacturing and Design Collaboration (eMDC) & Semiconductor Manufacturing (ISSM), Hsinchu, Taiwan, 1-4.
    6. Beemsterboer, B. , Land, M. , and Teunter, R. (2016), Hybrid MTO-MTS production planning: An explorative study, European Journal of Operational Research, 248(2), 453-461.
    7. Beemsterboer, B. , Land, M. , and Teunter, R. (2017a), Flexible lot sizing in hybrid make-to-order/make-tostock production planning, European Journal of Operational Research, 260(3), 1014-1023.
    8. Beemsterboer, B. , Land, M. , Teunter, R. , and Bokhorst, J. (2017b), Integrating make-to-order and make-tostock in job shop control, International Journal of Production Economics, 185, 1-10.
    9. Bemelmans, R. (1986), The Capacity-Aspect of Inventories, Springer, Heidelberg.
    10. Carr, S. , Gullu, A. , Jackson, P. , and Muckstadt, J. (1993), An exact analysis of a production-inventory strategy for industrial suppliers, Working Paper, School of Operations Research and Industrial Engineering, Cornell University, Ithaca, NY.
    11. Chang, S. H. , Pai, P. F. , Yuan, K. J. , Wang, B. C. , and Li, R. K. (2003), Heuristic PAC model for hybrid MTO and MTS production environment, International Journal of Production Economics, 85(3), 347-358.
    12. Fisher, M. , Ramdas, K. , and Ulrich, K. (1999), Component sharing in management of product variety: A study of automotive braking systems, Management Science, 45, 297-315.
    13. Forrester, J. (1961), Industrial Dynamics, MIT Press, Cambridge MA.
    14. Georgiadis, P. and Michaloudis, C. (2012), Real-time production planning and control system for job-shop manufacturing: A system dynamics analysis, European Journal of Operational Research, 216(1), 94-104.
    15. Ghalehkhondabi, I. , Ardjmand, E. , and Weckman, G. (2017), Integrated decision making model for pricing and locating the customer order decoupling point of a newsvendor supply chain, Opsearch, 54(2), 417-439.
    16. Ghalehkhondabi, I. , Sormaz, D. , and Weckman, G. (2016), Multiple customer order decoupling points within a hybrid MTS/MTO manufacturing supply chain with uncertain demands in two consecutive echelons, Opsearch, 53(4), 976-997.
    17. Gupta, D. and Benjaafar, S. (2004), Make-to-order, maketo- stock, or delay product differentiation? A common framework for modeling and analysis, IIE Transactions, 36(6), 529-546.
    18. Halawa, F. , Lee, I. G. , Shen, W. , Khan, M. E. , and Nagarur, N. (2017), The implementation of hybrid mtsmto as a promoter to lean-agile: A simulation case study for Miba Sinter Slovakia, Proceedings of the 2017 Industrial and Systems Engineers Conference, 1006-1011.
    19. Hax, A. and Meal, H. (1975), Hierarchical integration of production planning and scheduling. In: M. A. Geisler (ed.), TIMS Studies in Management Sciences, Logistics, Elsevier Publishing, Amsterdam, North Holland, 53-69.
    20. Hendry, L. and Kingsman, B. (1989), Production planning systems and their applicability to make-to-order companies, European Journal of Operations Research, 40(1), 1-15.
    21. Jia, Y. , Weng, W. , and Fujimura, S. (2017), A hybrid MTS-MTO production model with a dynamic decoupling point for flexible flow shops, Proceedings of the 16th International Conference on Computer and Information Science (ICIS), Wuhan, China, 803-807.
    22. Lee, C. and Chung, C. (2012), An inventory model for deterirating items in a supply chain with system dynamics analysis, Procedia-Social and Behavioral Sciences, 40, 41-51.
    23. Lee, H. (1996), Effective inventory and service management through product and process redesign, Operations Research, 44(1), 151-159.
    24. Lee, H. L. and Tang, C. S. (1997), Modeling the costs and benefites of delayed product differentiation, Management Science, 43(1), 40-53.
    25. Li, L. (1992), The role of inventory in delivery-time competition, Management Science, 38(2), 182-197.
    26. Meredith, J. and Aknic, U. (2007), Characterizing and structuring a new make-to-forecast production strategy, Journal of Operations Management, 25(3), 623-642.
    27. Nagib, A. N. M. , Adnan, A. N. , Ismail, A. , Halim, N. H. A. , and Khusaini, N. S. (2016), The role of hybrid make-to-stock (MTS)-make-to-order (MTO) and economic order quantity (EOQ) inventory control models in food and beverage processing industry. In: IOP Conference Series: Materials Science and Engineering (vol. 160, no. 1, p. 012003), IOP Publishing.
    28. Nguyen, V. (1998), A multiclass hybrid production center in heavy traffic, Operations Research, 46(3), S13-S25.
    29. Oğuz, C. , Salamn, F. , and Yalçın, Z. (2010), Order acceptance and scheduling decisions in make-to-order systems, International Journal of Production Economics, 125(1), 200-211
    30. Olhager, J. (2003), Strategic positioning of the order penetration point, International Journal of Production Economics, 85(3), 319-329.
    31. Poles, R. (2013), System dynamics modelling of a production and inventory system for remanufacturing to evaluate system improvement strategies, International Journal of Production Economics, 144(1), 189-199.
    32. Rabbani, M. and Dolatkhah, M. (2017), Simultaneous production planning of make-to-order (MTO) and make-to-stock (MTS) products using simulation optimization. Case study: Soren Restaurant, International Journal of Advanced Logistics, 6(1), 30-44.
    33. Rafiei, H. and Rabbani, M. (2012), Capacity coordination in hybrid make-to-stock/make-to-order production environments, International Journal of Production Research, 50(3), 773-789.
    34. Rafiei, H. , Rabbani, M. , and Alimardani, M. (2013), Novel bi-level hierarchical production planning in hybrid MTS/MTO production contexts, International Journal of Production Research, 51(5), 1331-1346.
    35. Rafiei, H. , Rabbani, M. , and Hosseini, S. H. (2014), Capacity coordination under demand uncertainty in a hybrid make-to-stock/make-to-order environment: A system dynamics approach, Scientia Iranica. Transaction E, Industrial Engineering, 21(6), 2315-2325.
    36. Rafiei, H. , Rabbani, M. , Vafa-Arani, H. , and Bodaghi, G. (2017), Production-inventory analysis of single-station parallel machine make-to-stock/make-to-order system with random demands and lead times, International Journal of Management Science and Engineering Management, 12(1), 33-44.
    37. Rajagopalan, S. (2002), Make-to-order or make-to-stock: Model and application, Management Science, 48(2), 241-256.
    38. Renna, P. (2016), Production control policies for a multistage serial system under MTO-MTS production environment, The International Journal of Advanced Manufacturing Technology, 83(1-4), 449-459.
    39. Samadhi, T. M. A. A. and Hoang, K. (1995), Shared computer-integrated manufacturing for various types of production environment, International Journal of Operations & Production Management, 15(5), 95-108.
    40. Soman, C. A. , Van Donk, D. P. , and Gaalham, G. (2004), Combined make-to-order and make-to-stock in a food production system, International Journal of Production Economics, 90(2), 223-235.
    41. Soman, C. A. , Van Donk, D. P. , and Gaalman, G. J. C. (2007), Capacitated planning and scheduling for combined make-to-order and make-to-stock production in the food industry: An illustrative case study, International Journal of Production Economics, 108(1-2), 191-199.
    42. Sridharan, S. V. (1998), Managing capacity in tightly constrained systems, International Journal of Production Economics, 56-57, 601-610.
    43. Sterman, J. D. (2000), Business Dynamics: Systems Thinking and Modeling for a Complex World, Irwin McGraw, New York, NY.
    44. Swaminathan, J. M. and Tayur, S. R. (1998), Managing broad product lines through delayed differentiation using vanilla boxes, Management Science, 44(12), S161-S173.
    45. Tsubone, H. , Ishikawa, Y. , and Yamamoto, H. (2002), Production planning system for a combination of make-to-stock and make-to-order products, International Journal of Production Research, 40(18), 4835-4851.
    46. Van Donk, D. P. (2001), Make to stock or make to order: The decoupling point in the food processing industries, International Journal of Production Economics, 69(3), 297-306.
    47. Williams, T. M. (1984), Special products and uncertainty in production/inventory systems, European Journal of Operations Research, 15(1), 46-54.
    48. Youssef, K. H. , Van Delft, C. , and Dallery, Y. (2004), Efficient scheduling rules in a combined make-tostock and make-to-order manufacturing system, Annals of Operations Research, 126(1-4), 103-134.
    49. Zaerpour, N. , Rabbani, M. , Gharehgozli, A. , and Tavakkoli- Moghaddam, R. (2008), Make-to-order or make-to-stock decision by a novel hybrid approach, Advanced Engineering Informatics, 22(2), 186-201.
    Do not open for a day Close