Resource-Based Cost Modeling - CyberLeninka

0 downloads 0 Views 341KB Size Report
of time because expected cost savings could not be realized. One reason for ... principles: source-specific, objective-oriented and valid. The approach ... Challenges in cost modeling of production networks. Even though ... It targets some aspects of the .... using comparison data from the statement of income and the values of ...
Available online at www.sciencedirect.com

ScienceDirect Procedia CIRP 63 (2017) 64 – 69

The 50th CIRP Conference on Manufacturing Systems

Resource-Based Cost Modeling – a New Perspective on Evaluating Global Production Networks Guenther Schuha, Jan-Philipp Protea, Torben Schmitza* a

Laboratory for Machine Tools and Production Engineering, Steinbachstraße 19, 52074 Aachen, Germany

* Corresponding author. Tel.: +49-241-80-28293. E-mail address: [email protected]

Abstract Production networks need to be constantly adapted so companies can handle the introduction of new products, tap the opportunities of new markets or significantly save costs. Especially cost reductions require an adequate way to predict the cost effects of network adaptions such as product allocations or location decisions. However, studies show that a high amount of network adaptions have been reversed after a short period of time because expected cost savings could not be realized. One reason for this is that simplistic cost estimations are still common in industrial practice since literature approaches are often too complex or require too much effort to apply them. The approach presented in this paper tackles the problem of the low applicability of existing cost models. It is based on a set of three guiding principles: source-specific, objective-oriented and valid. The approach is source-specific because it includes a resource-based cost modeling method that is used to evaluate the operational costs for each product within the network. Objective orientation is achieved by a flexible aggregation method that enables to define the level of detail necessary for a given decision situation. The validity of the cost model, a crucial element for decision makers, is created through a method that allows for calibrating and adjusting the model with standardized cost information. The result of the presented approach is a comparison of unit costs as well as a net present value calculation for a given decision situation. An exemplary application shows how the approach can be used as a support for designing the future production network of a company. © Published by by Elsevier B.V.B.V. This is an open access article under the CC BY-NC-ND license 2017The TheAuthors. Authors. Published Elsevier © 2017 (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the scientific committee of The 50th CIRP Conference on Manufacturing Systems. Peer-review under responsibility of the scientific committee of The 50th CIRP Conference on Manufacturing Systems Keywords: Cost Evaluation; Production Network; Global Networking

1. Introduction Markets for the majority of companies are spread around the world. Therefore, to follow their markets, companies produce more and more internationally. About 40% of the production capacities of all German companies with foreign production are located abroad [1]. This trend is emphasized by growing worldwide foreign direct investments (FDI) that have increased to about US$1.8 trillion by 2015 [2]. The main advantages of an internationalization of production and, hence, the creation of production networks, can be seen in strengthening a company’s position towards the competitors in terms of costs, gaining access to new markets and resources as well as reducing the distance to existing customers [3]. In the past, production networks have been redesigned and optimized individually. Nowadays, companies realize that a network is a

complex system whose performance can be enhanced by a systematic configuration [4]. A certain configuration or network alternative has effects on key performance indicators such as lead times, responsiveness, flexibility, financial performance or the cost structure [5]. The last indicator, the cost structure, is the focus of this paper. 2. Challenges in cost modeling of production networks Even though companies produce more and more abroad, the backshoring rate for production remains high. One key issue are expectations about possible cost savings that could not be realized [1]. Some of the main challenges are discussed in the following paragraphs. Cost-based Operations Research (OR) methods for designing and optimizing production networks have existed since the 1970s, but have not been introduced in producing

2212-8271 © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the scientific committee of The 50th CIRP Conference on Manufacturing Systems

doi:10.1016/j.procir.2017.03.354

Guenther Schuh et al. / Procedia CIRP 63 (2017) 64 – 69

companies to a high extent. Instead, simple investment appraisals are used [6,7]. A missing establishment of OR methods in industrial practice can be explained by low relevance, missing transparency and high efforts [8]. These deficits can be found in most advanced cost models. The allocation of costs has a high impact on the cost evaluation of production networks. Some costs such as overhead costs for indirect staff do not immediately depend on the production volume, but are caused by the product and/or process complexity [9]. A main deficiency of current cost models is the missing source-specific assignment of costs which leads to incorrect evaluations [10]. Many cost models for production networks do not include a systematic validation in order to reduce modeling efforts. This missing validation results in inadequate cost estimations for adapted network structures that cannot be achieved. To avoid this problem, validation by systematically calibrating and adjusting a cost model is necessary [11]. In summary, three categories of challenges for cost modeling of production networks can be described: x Low applicability due to high efforts and low transparency of existing cost models x Unsuitable source-specific assignment of costs x Missing validity due to unsuitable calibration and adjustment 3. State of the art In research, a number of approaches has been developed over time to improve cost modeling of production networks. Two main research fields can be distinguished that deal with the cost evaluation of production networks: OR-based network design and cost accounting for production networks. Approaches from both fields will be presented in the following paragraphs. Huebner develops an approach for strategic network design in the process industry. The planning process consists of three different phases: problem description, optimization of production networks and site selection. Regarding the above described three categories of challenges, the optimization approach of Huebner addresses mainly the process industry and has therefore a low applicability in other industries. It is not focused on source-specific assignment of costs. Validation is regarded, but not in a systematic way [7]. Schilling develops a planning and decision support for the management of production networks that is designed for the consumer goods industry and based on an optimization model. The approach includes the evaluation of the existing network including data validation as well as the generation of a objective configuration. It targets some aspects of the challenges since activity-based costing for overhead costs and a structured validation of the cost model is incorporated. Nevertheless, the general applicability of the approach can be considered low since it causes high efforts and is specifically designed for the consumer goods industry [9].

Schuh et al. develop a software tool for the configuration of production networks that includes an optimization model minimizing the total landed costs. The approach is filled with all necessary data for the application of the optimization model. This includes data regarding products, processes, location, as well as costs. The software tool allows for a detailed analysis of the optimized network structure. The approach of Schuh et al. can be applied to various industries but involves high efforts for data generation. A source-specific assignment of costs is addressed only to a low degree by integrating different characteristics of cost functions. Validation of data is not integrated in a systematic manner [12]. Further approaches for OR-based network design have been developed by e.g. Melo et al., Tsiakis and Papageorgiou or Yuan et al. but do not tackle the challenges more profoundly than the described examples [13,14,15]. LaLonde and Pohlen develop an activity-based cost accounting approach for performance measurement in supply chains. Their approach consists of six steps: First, key processes are identified. Second, the processes are broken down into activities. The next step is the identification of resources required to perform an activity. Step four comprises costing the activities by summing up the resource costs linked to an activity. The fifth step covers the definition of connections between activity costs and supply chain outputs. In the last step, the effects of cost drivers towards the costs of a supply chain are analyzed. The applicability and transparency of the approach of LaLonde and Pohlen are high, even though some efforts are necessary. The assignment of costs can be considered source-specific due to the implementation of an activity-based costing method. A validation of the approach is mentioned by the authors, but not described in detail [16]. The cost accounting model of Schulze et al. also evaluates the costs of supply chains based on activities (activity-based costing). The authors introduce a two-step methodology: the first step deals with the product design, the second addresses production and logistics. The core of the first step is a description of business activities. For those processes and activities, cost drivers are defined. By determining and varying the cost driver quantity, the cost effects of processes and activities can be analyzed. This is used for a selection of cost suppliers and a cost-efficient product design. In the second step, the design and optimization of the production network and the supply chain is targeted. By determining the standard time per activity as well as the cost per time unit, cost rates are determined which are used for the process cost calculation. The process costs are then utilized to evaluate relocations of activities or optimizations within the supply chain, e.g. by introducing a higher degree of automation. The approach of Schulze et al. is transparent and highly applicable. A sourcespecific assignment of costs is possible, even though different types of costs and their specifics are not regarded. A systematic way for validating the model is not included [17]. Further cost accounting approaches for production networks exist that have not been presented in detail. Examples are the approaches of Dekker and van Goor, as well as Pohlen and

65

66

Guenther Schuh et al. / Procedia CIRP 63 (2017) 64 – 69

Coleman [18,19]. Regarding the described challenges, only few new insights are presented. It can be stated that the described challenges have not been fully addressed by the existing state of the art. Therefore, a new methodology for the evaluation of production networks is proposed. 4. Methodology The methodology described in this chapter follows a different path than the state of the art. It is neither an OR approach nor a pure cost accounting approach. It is designed for the cost evaluation of defined production network structures. Three guiding principles represent the basis for the approach: the methodology should be source-specific, objective-oriented and valid. Those principles tackle the challenges described in Chapter 2. The approach should be source-specific, therefore allowing a cost evaluation that reveals interdependencies and creates transparency between key parameters of a production network. For this purpose, a resource-based cost model is developed. This modeling technique is more flexible than e.g. activity-based costing since the resource consumption and the cost evaluation of the consumption are modeled separately. The objective orientation enables to model cost for a specific purpose in a specific manner. Only the necessary elements to answer specific questions are included, e.g. the costs for a product before and after a relocation. Last, the validity of the model is a main feature. Methods for the calibration and adjustment are introduced. The purpose of those methods is for decision makers to gain trust in the model and to be able to finalize decisions based on the model’s outcome. The guiding principles are translated into a methodology that contains three different phases: the definition phase, the modeling phase and the evaluation phase. The three phases are illustrated in Fig. 1 and described in detail in the following subchapters.

I: Definition phase

II: Modeling phase

III: Evaluation phase

-

-

-

-

Definition of objectives and scope Determination of basic modelling parameters

-

Resource-based cost modeling Calibration and adjustment Preparation of cost projections

Unit cost comparison Discounted cash flow calculation

Fig. 1. Methodology for resource-based cost modeling of production networks.

4.1. Definition phase In the definition phase, the objective and the scope are defined. Furthermore, the alternatives to be evaluated are determined. Even though the cost evaluation plays the major role, other quantitative or qualitative objectives might be also important for a decision. Therefore, the importance of a cost evaluation in comparison to other criteria is assessed. This prioritization

can have an impact on the necessary level of detail of the cost model. With the scope, it is defined which products, locations, business functions, processes and types of costs are included in the model. The scope depends on the different network alternatives that should be evaluated. The structure of the model in terms of network, locations and functions within the scope can be described and visualized with a plane model such as the one developed by Westkaemper [20]. The regarded processes as well as the types of costs can be defined individually for a location/network. Reuter et al. present a proposal for processes and types of costs to be integrated in a cost model [21]. The description of the network alternatives is fundamental for all following steps. A useful support tool can be the illustration of the product structure, its processes and the alternatives for theses processes. An example for such an illustration can be found in Fig. 2. Legend:

Module mainboard (1x)

Mainboard placement

Alt. 1: Low-Cost Production

Wave soldering

Alt. 2: Know-How Protection in G

Alt. 1 China

Location

Alt. 2 Germany

Module case (1x) Punching

Forming

Alt. 1 Poland Alt. 2 Germany

Drilling

Preassembly

measurement device (52.000 pcs/year) Installation

Final assembly

Alt. 1 Germany …

Alt. 2 Germany

Fig. 2. Illustration of two network alternatives including the product and production structure.

4.2. Modeling phase The modeling phase is the main phase of the new methodology. The first step is the creation of a resource-based cost model. Step two is the calibration and adjustment of the created model. Finally, in step three, cost projections for alternative structures are prepared. Before the resource-based cost model can be created, the available cost data have to be aggregated. The goal of the cost model is not absolute accuracy. The idea behind the aggregation (and also the calibration and adjustment) is to be as detailed as necessary. For this purpose, the level of detail of the products, the functions as well as the processes within the model have to fit to the decision situation. The principle “from rough to detail” is followed for all products, functions and processes. If for the decision situation it is sufficient to look at a product class or location as a whole, further differentiation is not needed. In this way, the minimum level of aggregation for the cost model is determined. The resource-based cost model for production networks is based on the research of Schuh for the cost evaluation of product variants [22]. The idea of the cost model is to assess the consumption of resources and the costs of the consumed resources separately. For this purpose, two functions, a cost function as well as a consumption function, are developed. The

67

Guenther Schuh et al. / Procedia CIRP 63 (2017) 64 – 69

Consumption of resources (e.g. raw material volume)

Cost function

Costs (e.g. raw material costs)

Consumption function

Cost driver (e.g. production volume)

Fig. 3: Cost function and consumption function of the resource-based cost model [21].

The resources and cost drivers included in the model need to be determined specifically for a certain company and/or network. Possible resources can be staff, materials, energy, machinery, equipment and tools, factory and office equipment, buildings and transport. The cost drivers represent the production program, therefore they can be divided into quantity, number of variants and number of production orders per product. The relationships between cost drivers, resources as well as costs of a certain type are described by the cost function and the consumption function. For both functions, it is possible to use different courses. Typically, a selection of linear, declining and fixed step courses for the consumption function, as well as linear and declining courses for the cost function can cover most resources/cost types. More information about the cost and the consumption function of the cost model can be found in Reuter et al. [21,23]. The foundation of any cost evaluation is the creation of a current state model. For this purpose, consumption and cost functions for all cost types and products of the current state have to be determined. This first quantification should be done workshop-based with a team of experts from controlling and production. First, the team creates general hypotheses for all parameters of all functions within the cost model. Afterwards, these hypotheses for the parameters are adapted for each specific product. Data from business-software (e.g. ERP, PLM, etc.) can be used to confirm the hypotheses. This path is followed because a direct integration of data from business software is often not possible due to the necessary aggregation. Defining hypotheses and backing them with expert knowledge and data reduces the effort for creating the model. Furthermore, the following calibration and adjustment of the model allows for the correction of mistakes that may occur in this step.

The calibration and adjustment of the current state cost model enables the generation of the level of accuracy needed for a given decision situation. Depending on the available cost data, the calibration can be done data-based or experiencebased. The main requirement is that the input used for the calibration is not directly linked to the input used in the current state model. If reliable data is available, a data-based approach should be generally preferred and used for calibration and adjustment purposes. For the calibration, standard values are compared to the output values of the cost model. Two different kinds of deviations are possible: the model values can be either higher than the standard value or lower than the standard value. Fig. 4 demonstrates the calibration of salaries for production control using comparison data from the statement of income and the values of a cost model. The calibration only determines a difference between a model and a standard value. The adjustment allows for adaptions of the model and therefore a better fit. An adjustment can be done by changing the parameters of the consumption function, the cost function or, if a sufficient level of detail cannot be generated, by adapting the level of aggregation of products, processes or functions. For the adjustment, a sufficient level of accuracy needs to be defined for each consumed resource and each type of cost. One way to achieve this is the construction of a required level of accuracy that can be compared to the actual level of accuracy defined by the sum of values of the cost model divided by the comparison value. The required level of accuracy is based on weighted, qualitative criteria such as decision relevance, consistency or sensitivity of the cost type. A similar approach for the assessment of the value of information in decision situations can be found in Mueller [24]. Raw data e.g. statement of income

Prepared standard values

Salaries

Salaries Production Control

Σ Values of the cost model Salaries PC Product group 1 Salaries PC Product group 2

Deviation standard-model

relation between the two functions can be illustrated using a nomogram (see Fig. 3). For the purpose of evaluating production networks, the original model of Schuh is adapted so that it is based on the production program of the network. Each product is modeled individually for each function or process as well as each type of cost using so called cost drivers. This allows for a high flexibility and adaptability of the model. At the same time, the modeling approach creates transparency in terms of interdependencies between products, necessary resources and the resulting costs.



Fig. 4: Comparison of standardized data (statement of income) and the resource-based cost model.

The adjusted model of the current state is the basis for future cost estimations. Nevertheless, more data is necessary to prepare such estimations. The cost model needs to be adapted for new decision alternatives that are not covered by the current state. Those can be new products, new materials or new processes. These new decision alternatives lead to new or adapted consumption and cost functions and therefore require new data that have to be made available for the cost model. As often as possible, data should be derived from comparable products or processes or available data, e.g. from machine manufacturers. If no data is available, expert estimations or external databases have to be used. Not only do product or process parameters change, also market conditions and

68

Guenther Schuh et al. / Procedia CIRP 63 (2017) 64 – 69

therefore the prices of resources change. Thus, for all points of time in the future for which a cost estimation is needed, the market conditions have to be described. The most pragmatic way is to derive the future conditions from annual percentage changes of the current market conditions. This leads to adaption of the parameters of each cost function. 4.3. Evaluation phase In the evaluation phase, the output of the resource-based cost model is presented so that decisions can be derived. The results are presented in two different ways: x Unit cost comparison x Discounted cash flow calculation

cases (equal to the quantity of measurement devices) and the selected resource is the weight of the material (mainly stainless steel). The data for the current state has been calibrated and adjusted with comparison data based on expert knowledge. For both the consumption as well as the cost function, a linear course was selected. It is assumed that the consumed materials per piece (0.0007 tonnes/piece) will decrease by 16% annually due to improved processes and a lower scrap rate. Furthermore, the material cost rate (550 €/tonne) is assumed to be growing by about 18% per year. Therefore, based on the functions depicted in Fig. 5, material costs are projected to increase from 20,020 € to 22,700 € with a corresponding increase of the cost driver (quantity) from 52,000 pieces to 60,000 pieces. Alternative 1: Low-Cost Production

Location: Poland Consumed material (steel) (weight [tonne])

2016

With the unit cost comparison, it can be stated whether one network alternative (e.g. the current one) results in higher or lower unit costs than one or more other alternatives. Also, it can be determined if the unit costs fulfill requirements that have been defined by decision makers. If the requirements are not fulfilled, the efforts to create a discounted cash flow calculation are not necessary. For the unit cost comparison, the costs of a product along all processes at all production plants have to be included. Hence, the visualization of the product structure in Fig. 2 can be helpful. Using the resource-based cost model, the sum of all (operative) costs of a product can be determined. This value needs to be divided by the quantity of the final product to get to the average unit costs. The discounted cash flow calculation enables to evaluate the net present value of an adaption of a network structure. The cash flow of each period within a defined timespan is discounted with an interest rate to receive the present value. Only those network alternatives that have a positive sum of present values along the time span (net present value) are preferable. For the discounted cash flow calculation, only cashbased costs are taken into account. Therefore, possible imputed costs such as depreciations that might have been integrated in the resource-based cost model have to be removed. Instead, investments, ramp-up and restructuring costs that arise within the defined time span have to be integrated. 5. Exemplary application The developed methodology is exemplarily applied using data from a project where the relocation of a measurement device was to be decided (compare Fig. 2). Two alternatives are analyzed. The first alternative is the current decentralized low-cost production to achieve low unit costs, the second alternative aims to concentrate the know-how centrally by producing everything near the headquarter. The company is willing to accept unit costs up to 10% higher in 2018 if their intellectual property can be protected. For confidentiality reasons, the data have been modified with factors. Fig. 5 illustrates the example of a nomogram for the material costs for the case production in Poland in 2016 and the following 2 years. The selected cost driver is the quantity of

2017

2016

36,4

2017

31,8

2018

2018 29,6

22,700 20,020 20,610 Material costs (steel) (costs [€])

52,000 54,000

60,000 Cost driver (quantity [piece])

Cost function:

Consumption function:

mc m, md, P, t = (1,18)t-t0 x 550 x crm, md, P, t

crm, md, P, t = (0,84)t-t0 x 0.0007 x dm, md, P

Legend: to Current point of time (2016) t point of time

dm, md, P, t Cost driver for material m and product measurement device md (case) at location Poland at point of time t crm, md, P, t Consumed resource (material) for product measurement device md (case) at location Poland at point of time t mc m, md, P, t Total material costs for product measurement device md (case) at location Poland at point of time t

Fig. 5: Nomogram of material costs for the case of the product measurement device at location Poland (Alternative 1).

Cost and consumption function as shown in the given example have been created for all functions, processes and cost types at each location that is involved in the production of the measurement device described in Fig. 2. With all functions at hand, the unit cost comparison can be build up. For this purpose, all cost types are summed up and divided by the quantity of the final product. In Fig. 6, the unit cost comparison for the production of the measurement device with alternative 1 and alternative 2 in 2018 is shown. The material costs for the case for each final product from the example in Fig. 5 are illustrated in detail. It can be seen that the unit costs are about 15% higher for alternative 2 than for alternative 1. Therefore, alternative 2 is rejected and a discounted cash flow calculation is not pursued. 5,78 €/unit

6,62 €/unit

wage costs Example

Material costs case

material costs case other material costs

Total: 22,700 € quantity: 60000 units unit cost: 0,38 €/unit Alternative 1 2018

labor costs

+15%

other costs

Alternative 2 2018

Fig. 6: Unit cost comparison of product measurement device.

Guenther Schuh et al. / Procedia CIRP 63 (2017) 64 – 69

Some of the key features of resource-based cost modeling of production networks could be demonstrated in this example. The flexible, modular structure in terms of scope and level of detail allows for a reduction of efforts to a necessary minimum. Transparency towards interdependencies is created by separate modeling of the resource consumption and the cost evaluation of the consumed resources. This allows for modeling changes in internal conditions (e.g. improved processes) and external conditions (e.g. increasing prices) separately. The example only shows a first application. For a full verification, a study was initiated in which the methodology will be systematically verified. This study is conducted in cooperation with a first-tier automotive supplier company. The goal of the verification study is to assess whether the developed methodology improves the challenges listed at the end of Chapter 2. The challenges in terms of effort and transparency are assessed using expert interviews in which application experiences are gathered and analyzed. Furthermore, it is planned to verify the new cost assignment technique by modeling network adaptions from the past and compare the costs projected by the model to the accrued costs after the network adaption. Finally, the calibration and adjustment procedure proposed in this paper will be applied and it is demonstrated which level of validity can be achieved. First results of the validation study are expected to be available soon. These results will be part of the following publication about the resource-based cost model for production networks. 6. Conclusion Cost evaluation for the purpose of redesigning production networks is challenging. Therefore, in this paper, a new methodology for resource-based cost modeling of production networks that enables the creation of a source-specific, objective-oriented and valid cost model is presented. The exemplary application demonstrated some of the main benefits that can be achieved with this approach. The modularity of the approach allows for further integration of specific evaluations. It is planned to add methods to evaluate the costs of external uncertainties as well as remanence costs. Further expansions are in discussion. Acknowledgements The authors would like to thank the German Research Foundation (DFG) for funding this work in the research and development project "Cluster of Excellence – Integrative Production Technology for High Wage Countries”. References [1] Zanker C, Kinkel S, Maloča S. Globale Produktion von einer starken Heimatbasis aus: Verlagerungsaktivitäten deutscher Unternehmen auf dem Tiefstand. Mitteilungen aus der ISI-Erhebung Modernisierung der Produktion 2013;63:1-12. [2] UNCTAD. World Investment Report 2016: Investor Nationality – Policy Challenges. New York: United Nations; 2016.

[3] Jacob F. Quantitative Optimierung dynamischer Produktionsnetzwerke (Doctoral dissertation). Aachen: Shaker; 2006. [4] Friedli T, Mundt A, Thomas S. Strategic management of global manufacturing networks: Aligning strategy, configuration, and coordination. Berlin, Heidelberg: Springer; 2014. [5] Hayes R, Pisano G, Upton D, Wheelwright S. Operations, strategy, and technology: Pursuing the competitive edge. Hoboken, NJ: Wiley; 2005. [6] Vos B, Akkermans H. Capturing the dynamics of facility allocation. International Journal of Operations & Production Management 1996;16(11):57-70. [7] Huebner R. Strategic Supply Chain Management in Process Industries: An Application to Specialty Chemicals Production Network Design. Berlin, Heidelberg: Springer; 2007. [8] Meyer T. Globale Produktionsnetzwerke: Ein Modell zur kostenoptimierten Standortwahl (Doctoral dissertation). Aachen: Shaker; 2006. [9] Schilling R. Manufacturing network development in fast-moving consumer goods industries (Doctoral dissertation). Hamburg: Kovač; 2012. [10]Gunasekaran A, Williams HJ, McGaughey RE. Performance measurement and costing system in new enterprise. Technovation 2005;25(5):523533. [11]Meyer T, Jacob F. Network design. In: Abele E. et al. (Eds.). Global Production: A Handbook for Strategy and Implementation. Berlin, Heidelberg: Springer; 2008. p. 140-190. [12]Schuh G, Potente T, Kupke D, Varandani R, Hausberg C. An Evolutionary Approach for Global Production Network Optimisation. Procedia CIRP 3 2012;3:382–387. [13]Melo MT, Nickel S, Saldanha da Gama F. Dynamic multi-commodity capacitated facility location: A mathematical modeling framework for strategic supply chain planning. Computers & Operations Research 2006;33(1):181–208. [14]Tsiakis P, Papageorgiou LG. Optimal production allocation and distribution supply chain networks. International Journal of Production Economics 2008;111(2):468–483. [15]Yuan XM, Low JMW, Yeo WM. A network prototype for integrated production-distribution planning with non-multi-functional plants. International Journal of Production Research 2012;50(4):1097–1113. [16]LaLonde BJ, Pohlen TL. Issues in Supply Chain Costing. International Journal of Logistics Management 1996;7(1):1–12. [17]Schulze M, Seuring S, Ewering C. Applying activity-based costing in a supply chain environment. International Journal of Production Economics 2012;135(2):716–725. [18]Dekker HC, van Goor AR. Supply Chain Management and Management Accounting: A Case Study of Activity-Based Costing. International Journal of Logistics Research and Applications 2010;3(1):41–52. [19]Pohlen TL, Coleman BJ. Evaluating internal operations and supply chain performance using EVA and ABC. SAM Advanced Management Journal 2005;70(2):45. [20]Westkaemper E. Einführung in die Organisation der Produktion. Berlin, Heidelberg: Springer; 2006. [21]Reuter C, Prote J, Schmitz T. A top-down/bottom-up approach for modeling costs of a manufacturing network. In: Proceedings 23rd EurOMA Conference - Interactions. Trondheim, Norway. June 17th-22nd 2016. Trondheim, Norwegian University of Science and Technology (NTNU); 2016. p. 1-10. [22]Schuh G. Gestaltung und Bewertung von Produktvarianten: Ein Beitrag zur systematischen Planung von Serienprodukten (Doctoral dissertation). Düsseldorf : VDI-Verlag; 1989. [23]Reuter C, Prote J, Schmitz T. Cost modelling approach for the source specific evaluation of alternative manufacturing networks. In: Proceedings APMS 2016 Advances in Production Management Systems - Production Management Initiatives for a Sustainable World, Iguassu Falls, Brazil. September 3rd-7th, 2016. Umarizal, Universidade Paulista (UNIP): 2016; p. 1-8. [24]Mueller A. Informationsbeschaffung in Entscheidungssituationen (Doctoral dissertation). Ludwigsburg: Verlag Wissenschaft&Praxis; 1991

69