Simultaneous investment, operations, and financial planning in supply chains: A value-based optimization approach

Int. J. Production Economics 140 (2012) 559–569

Contents lists available at SciVerse ScienceDirect

Int. J. Production Economics journal homepage: www.elsevier.com/locate/ijpe

Simultaneous investment, operations, and financial planning in supply chains: A value-based optimization approach G.J. Hahn, H. Kuhn n Department of Supply Chain Management & Operations, Catholic University of Eichstaett-Ingolstadt, Auf der Schanz 49, 85049 Ingolstadt, Germany

a r t i c l e i n f o

abstract

Article history: Received 14 April 2010 Accepted 14 February 2012 Available online 27 February 2012

Asset utilization is a major mid-term lever to increase shareholder value creation. Since rough-cut planning of capacity (dis-)investments is performed at the long-term level, detailed timing of adjustments remains for the mid-term level. In combination with capacity control measures, capacity adjustment timing can be used to optimize asset utilization. This paper provides a corresponding framework for value-based performance and risk optimization in supply chains covering investment, operations, and financial planning simultaneously. We illustrate the benefits of the approach using a case-oriented example, and highlight the value of using flexible capacity options and postponing of capacity-related decisions in an uncertain environment. & 2012 Elsevier B.V. All rights reserved.

Keywords: Capacity planning Sales and operations planning Robust optimization Value-based management Performance and risk management

1. Introduction Since creating shareholder value is commonly considered the paramount business goal (Young and O’Byrne, 2001), frameworks for value-based management (VBM) are also discussed within the supply chain context (Walters, 1999; Lambert and Pohlen, 2001). Top-level performance metrics such as discounted Free Cash Flow (FCF) or Economic Value Added (EVA) and corresponding value driver trees to drill down the performance metric into operational levers are prevalent concepts of VBM (Rappaport, 1998). Risk implications are typically considered indirectly via risk-adjusted cost of capital (Kaplan and Atkinson, 1998). In contrast to the aforementioned explanatory frameworks, Lainez et al. (2009) and Hahn and Kuhn (2011b) provide model-driven approaches to value-based performance and risk management in supply chains. Whilst Lainez et al. (2009) focus on the long-term level of strategic network design for a planning period of 2–10 years, Hahn and Kuhn (2011b) cover the mid-term level of sales and operations planning with a planning period of 6–18 months (Fleischmann et al., 2008). At the mid-term level, asset utilization is one of the major value drivers from a value-based planning perspective besides operating profit margin and operational cash flow (Walters, 1999). Capacity (dis-)investments in technical equipment and capacity control measures modifying supply and/or demand represent the two levers to manage asset utilization (Olhager et al., 2001; Buxey, 2003). Hahn

n

Corresponding author. Tel.: þ49 841 937 1820; fax: þ 49 841 937 1955. E-mail addresses: [email protected] (G.J. Hahn), [email protected] (H. Kuhn). 0925-5273/$ - see front matter & 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpe.2012.02.018

and Kuhn (2011b) only focus on capacity control measures, and do not consider capacity (dis-)investments. Capacity adjustments can create additional value, but involve considerable risk potential due to costs of overcapacity or lost sales as well as physical degradation and depreciation (van Mieghem, 2003). Moreover, physical (dis-)investment decisions are inextricably interlinked with the corresponding financial decisions (Shapiro, 2007) and their impact on liquidity as well as overall value creation. An integrated approach to simultaneous investment, operations, and financial planning is therefore required that considers value-based implications. Capacity adjustments and equipment replacement typically involve a planning period of several years depending on the average useful life of the machine, and NPV-based approaches are thus utilized to evaluate the investment decision (Luss, 1982). Corresponding decisions are considered together with decisions on facility locations at the long-term level of strategic network design (Goetschalckx and Fleischmann, 2008). However, decision models for strategic network design only provide support on sizing and rough-cut timing of capacity (dis-)investments due to their longterm perspective and aggregated (semi-)annual time buckets (Fleischmann et al., 2008). Detailed timing of capacity adjustments and equipment replacement remains for the mid-term level. Consequently, an integrated approach to capacity (dis-)investment timing and capacity control as part of sales and operations planning (S&OP) is required to manage asset utilization comprehensively. A corresponding unified framework has not yet been discussed, especially with respect to robust and risk-mitigating strategies in capacity (dis-)investment planning. The aim of this paper is to develop a decision support framework for mid-term investment, operations, and financial planning in supply chains utilizing an integrated approach to value-based

560

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

performance and risk optimization. We extend the paper of Hahn and Kuhn (2011b) to develop a comprehensive approach to capacity management taking into account related (dis-)investment and financing decisions from a value-based perspective. The remainder of this paper is structured as follows: Section 2 provides a literature review on the domains relevant for this research. In Sections 3 and 4, we outline the conceptual approach and describe a corresponding decision model. Section 5 highlights implications of the approach using a case-oriented example. We conclude the paper in Section 6 with a summary of the findings and an outlook for further research.

2. Literature review Following the outline of the article, the literature review covers four domains relevant for the problem in focus: (i) valuebased performance and risk optimization, (ii) integrated capacity (dis-)investment and financial planning, (iii) integrated capacity and operations planning in supply chains, and (iv) robust capacity planning under uncertainty. Value-based performance and risk optimization. Concepts and metrics for supply chain performance management are widely discussed in the pertinent literature (Kleijnen and Smits, 2004; Gunasekaran et al., 2004; Cai et al., 2009). Corresponding frameworks such as the Supply Chain Operations Reference (SCOR) model have been genuinely developed for supply chain management (Supply Chain Council, 2010) or are adapted from general management literature such as the balanced scorecard and activity-based costing (Liberatore and Miller, 1998). However, the aforementioned frameworks have two major drawbacks from a decision support perspective (Cai et al., 2009): first, they cover a multitude of different metrics, but do not propose a paramount performance metric; second, they omit interdependencies and trade-offs between the metrics, and do not provide insight into cause-and-effect relationships. For example, the SCOR model covers five coequal but non-comprehensive as well as interdependent and partially conflicting top-level financial metrics. In contrast, value-based approaches apply one paramount and comprehensive performance metric (Young and O’Byrne, 2001) that can be used to consistently manage all (dis-)investment, operations, and financial decisions for value creation. Value-based approaches in supply chain management have received increasing attention since the early work of Christopher and Ryals (1999) investigating supply chain strategy and its impact on shareholder value creation. Walters (1999) and Lambert and Pohlen (2001) develop EVA-based value driver trees to relate operational supply chain performance levers to overall value creation. In contrast to these explanatory frameworks, model-driven approaches to value-based performance and risk management in supply chains are provided in Lainez et al. (2009) and Hahn and Kuhn (2011b). Lainez et al. (2009) focus on the long-term level of strategic investment and financial management, optimizing shareholder value according to the discounted Free Cash Flow method. Option contracts are utilized to manage risk in supplier–customer relationships. Hahn and Kuhn (2011b) cover mid-term sales, operations, and working capital management and implement an EVA-based objective function. A direct approach to risk management is applied using downside riskbased metrics and scenario-based robust optimization methods. However, they do not consider aspects of capacity (dis-)investment timing to bridge the gap between the long-term and midterm planning levels. Integrated capacity (dis-)investment and financial planning. In an early paper, Luss (1982) provides a comprehensive literature survey on decision models for capacity expansion and equipment

replacement. Recent literature reviews in this field covering a broad range of industries and different methodological approaches can be found in van Mieghem (2003), Wu et al. (2005), and Julka et al. (2007). More qualitative approaches to investment decision-making are described in Pirttila¨ and ¨ Sandstrom (1995), Olhager et al. (2001), and Ojala and Hallikas ¨ (2006). Pirttila¨ and Sandstrom (1995) integrate the capital budgeting process of a company with manufacturing strategy to comprehensively manage a portfolio of individual investment decisions. Olhager et al. (2001) provide a framework for longterm capacity (dis-)investment management linking manufacturing strategy and S&OP. Ojala and Hallikas (2006) investigate investment decisions under uncertainty and risk in buyer-dominating supplier networks. However, implications for integrated capacity (dis-)investment and financial management have not been discussed so far. Capacity (dis-)investment planning covers four major decision problems: location, technology, sizing, and timing (Luss, 1982). Although capacity reductions involve the same considerations as capacity expansions (Olhager et al., 2001), disinvestment decisions are only covered marginally in respective models (Luss, 1982). Location, size, and timing of capacity expansions/reductions are typically included in decision models for strategic network design (Goetschalckx and Fleischmann, 2008). Aspects of technology selection and equipment replacement are covered separately from capacity adjustment planning (Li and Tirupati, 1994; Rajagopalan, 1998). However, considerations with respect to economies of scale and optimal operating value require an integrated approach (van Mieghem, 2003). Moreover, financial implications regarding accounting policies, capital budgeting, and costs of invested capital need to be considered (Julka et al., 2007). Majumdar and Chattopadhyay (1999) and Lavaja et al. (2006) develop decision models for integrated capacity (dis-)investment and financial planning in power systems as well as supply chains in the process industry. However, aspects of integrated capacity adjustment and financial planning at the mid-term level are not covered due to the long-term perspective. Integrated capacity (dis-)investment and operations planning. Bradley and Arntzen (1999) and Rajagopalan and Swaminathan (2001) investigate integrated mid-term production and capacity expansion planning to analyze the trade-off between capacity and inventories for different demand patterns. Bradley and Arntzen (1999) implement EVA as a value-based performance metric and examine two case studies with seasonal demand patterns. However, they omit financial flows and their impact on economic value creation. Rajagopalan and Swaminathan (2001) analyze discrete capacity acquisitions in an environment with gradual demand growth resulting in excess capacity in the period subsequent to the investment. Bhutta et al. (2003) and Hsu and Li (2009) investigate integrated capacity and supply chain operations planning at the long-term level. Bhutta et al. (2003) consider a multi-national company and analyze exogenous factors such as exchange and tariff rates. Hsu and Li (2009) examine an example from the semiconductor industry and investigate optimal supply chain network design incorporating economies of scale. A unified approach for capacity adjustment and supply chain operations planning at a mid-term level covering both capacity adjustment and equipment replacement has not yet been discussed. Robust capacity planning under uncertainty. A variety of papers discuss decision models for capacity planning under uncertainty. Eppen et al. (1989) and Paraskevopoulos et al. (1991) consider different levels of risk aversion and evaluate sensitivities to implement more robust solutions. Bok et al. (1998) and Aghezzaf (2005) present robust optimization approaches for capacity expansion and facility location planning in supply chains based on the robustness concepts of Mulvey et al. (1995). Barbaro

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

and Bagajewicz (2004) investigate different risk metrics and risk management methods in capacity expansion planning utilizing two-stage stochastic programming. Poojari et al. (2008) use a two-stage robust optimization approach to determine strategic decisions on locations, technology selection, and capacity adjustment policies. Robustness and risk impact are evaluated based on different risk metrics utilizing a simulation-based approach. Barahona et al. (2005) and Rastogi et al. (2011) investigate robust approaches to capacity planning in the semiconductor industry. Although robust optimization involves risk considerations, a unified approach to risk management and robust optimization in capacity (dis-)investment planning has not yet been developed. Ahmed et al. (2003) examine a multi-resource capacity expansion problem under uncertainty utilizing a multi-stage stochastic programming approach. Applications of multi-stage stochastic programming to the semiconductor industry are provided in Chen et al. (2002) and Huang and Ahmed (2009). Chen et al. (2002) investigate the interdependencies of technology selection and capacity planning. Huang and Ahmed (2009) focus on the value of a multi-stage approach allowing for the revision of investment decisions. Since multi-stage models require a high amount of accurate and detailed information in a real-life planning situation, sequential or rolling horizon approaches are typically applied (Scholl, 2001). Gupta and Rosenhead (1968) develop a robust sequential approach to postpone and revise capacity-related decisions. However, corresponding concepts have not yet been adapted in the context of two-stage stochastic programming. In summary, a comprehensive model-driven approach for midterm investment, operations, and financial management has not yet been presented to support monthly business planning in a VBM context. Although capacity management represents an important mid-term value lever especially in equipment-intensive industries, decision models for integrated capacity adjustment and capacity control management improving asset utilization and reducing lost sales have not yet been discussed. Capacity-related decisions and decisions on mid-term financing can involve substantial risk potential. Common risk management approaches typically use indirect methods such as risk premiums or apply sophisticated multi-stage stochastic programming approaches with complex scenario trees. However, a pragmatic approach to derive robust and risk-mitigating strategies in a typically sequential decision-making process has not yet been described. We thus develop a comprehensive mid-term approach for value-based performance and risk management in supply chains integrating investment, operations, and financial planning.

3. Conceptual approach 3.1. Overview A comprehensive framework for mid-term investment, operations, and financial planning in supply chains is developed in this section to support a VBM approach in monthly business planning. EVA as a prevalent performance metric of VBM is implemented due to the high intelligibility and practicability of the concept (Young and O’Byrne, 2001). Robust optimization methods are applied for risk management explicitly considering the riskaverse preference of business decision-makers by implementing different robustness criteria (Scholl, 2001). Since an integrated approach to performance and risk management is required to optimize shareholder value comprehensively (Rappaport, 1998; Ritchie and Brindley, 2007), we thus build on the decision framework of Hahn and Kuhn (2011b).

561

The decision model pursues a mixed-integer linear programming approach for mid-term S&OP in a company supply chain. Capacity targets from strategic network planning serve as instructions for detailed timing of capacity adjustments and equipment replacement. Aspects of mid-term financial management are integrated to ensure required funding for investment activities. A general modeling approach is applied to cover different depreciation methods and capital loss of technical equipment. The robust optimization approach is developed further to allow for partial postponement and revision of (dis-)investment and financing decisions supporting a pragmatic sequential decision-making process as described in Gupta and Rosenhead (1968). 3.2. Value-based performance and risk optimization An integrated approach for value-based performance and risk optimization is implemented in the objective function based on the EVA concept. EVA determines economic profit, i.e., operating profit minus total costs of invested capital (Young and O’Byrne, 2001). In (1), EVA in period t equals net operating profit after tax NOPAT minus the capital charge (Kaplan and Atkinson, 1998). The capital charge on invested capital is derived from net operating assets NOA at the end of the previous period t1 and weighted average cost of capital iwacc (Kaplan and Atkinson, 1998). Performance metrics such as EVA cannot be influenced at the top level and are thus drilled down into separate value drivers (Rappaport, 1998) as illustrated in Fig. 1. EVAt ¼ NOPATt NOAt1  iwacc

ð1Þ

Operating profit margin, asset utilization, and operational cash flow can be considered the major value drivers from a mid-term planning perspective (Walters, 1999). The EVA concept covers each of the five performance metrics presented in the SCOR model (Supply Chain Council, 2010). The EVA categories costs of goods sold (COGS) and expenses cover the cost-related metrics ‘costs of goods sold’ and ‘supply chain management costs’ of the SCOR model. The capital charge on net assets in the EVA concept corresponds to the asset-related metrics ‘return on supply chain fixed assets’, ‘return on working capital’ and ‘cash-to-cash cycle time’ of the SCOR model. In summary, optimizing EVA results in an optimal trade-off between the five performance metrics of the SCOR model. Value drivers depend on performance levers and operational risk occurrence. Performance levers cover the decision variables of mid-term sales, operations, and working capital management (Hahn and Kuhn, 2011a) as well as capacity (dis-)investment and debt management. Operational risks result from the uncertainty of future events in the ordinary course of business as opposed to disruption risks from natural or man-made disasters (Tang, 2006). In our approach, we focus on risk due to demand uncertainty as an external factor and use a scenario-based robust optimization approach to develop contingency plans for risk impact mitigation (Mulvey et al., 1995). We assume that internal risks such as machine breakdowns or quality defects can be addressed by improved maintenance and quality assurance. Robust optimization (RO) represents a generalization of stochastic programming explicitly considering the risk-averse preference of the decision-maker (Mulvey et al., 1995). RO thus aims at deriving plans that are sufficiently insensitive to the influence of imperfect information (Scholl, 2001). Three different robustness criteria are implemented in the decision model according to Scholl (2001): model, solution, and objective robustness. We utilize a total model-robust ‘fat solution’ design to provide decisive and feasible solutions for each scenario. The decisionmaker balances solution and objective robustness based on the individual risk preference (Hahn and Kuhn, 2011b). Whilst

562

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

Fig. 1. A decision-oriented value driver tree of the EVA concept.

Fig. 2. Archetypes of capacity (dis-)investment timing (based on Olhager et al., 2001).

solution robustness represents a risk-neutral preference seeking for optimal results, a risk-averse decision-maker aims at objective robustness to attain a predefined aspiration level (Scholl, 2001). The aspiration level corresponds to the desired outcome from which the adverse deviation is calculated to quantify risk impact (Vaughan, 1997). X X FðgÞ ¼ g  pr s  EVAs ð1gÞ  pr s  maxf0; AEVAs g ð2Þ sAS

|fflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflffl} expected EVA

sAS

|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} downside risk of EVA

In (2), we use the EVA concept with its implicit aspiration level of 0, i.e., that total costs of invested capital are covered, to develop an objective function F for integrated value-based performance and risk optimization (Hahn and Kuhn, 2011b). The performance or level component corresponds to expected EVA. Downside risk of EVA defined as the first-order lower partial moment (Nawrocki, 1999) represents the risk or deviation component with the aspiration level A ¼0 as the desired outcome. pr denotes the probability of scenario s in the scenario set S. The decision-maker balances performance and risk according to the risk preference parameter g A ð0; 1 covering the full range of weakly risk-averse preferences (Scholl, 2001). For g ¼ 1, the objective function results in the limiting case of a risk-neutral expected value criterion obtaining relative solution robustness. g-0 represents a highly risk-averse decision-maker mainly focusing on objective robustness with respect to the aspiration level implemented in the downside risk metric. 3.3. Robust capacity adjustment and mid-term financial planning Capacity adjustment planning involves different subproblems as discussed above. We focus on detailed timing of capacity

adjustments and equipment replacement at the mid-term level. Three different archetypes of capacity adjustment timing can be distinguished (Olhager et al., 2001): lagging, tracking, and leading (see Fig. 2). In the leading approach, capacity supply always exceeds capacity demand and provides high flexibility as well as reliable lead times at the cost of idle capacity. By contrast, in a lagging approach capacity demand always exceeds capacity supply to ensure high asset utilization, resulting in considerable lost sales or long lead times. A tracking approach represents a compromise between the two pure strategies, minimizing the deviation of capacity supply and demand. In our framework, we do not predetermine a specific timing strategy, but optimize the trade-off between lost sales and costs of idle capacity. We exclude the decision problem of technology selection and consider one type of technical equipment with a deterministic useful life and constant capacity. Depreciations and the accounting value of technical equipment are calculated based on a predefined series of book values according to common depreciation methods and a salvage value at the end of useful life (Nobes and Parker, 2008). A time-independent capital loss factor for disinvestments is used to account for a typically lower market value compared to book value, assuming that no further adjustment costs arise. Since the physical and financial perspectives of business are inextricably interlinked (Shapiro, 2007), sufficient (debt) funding has to be provided for capacity investments. To ensure duration congruence in the balance sheet, we consider mid- to long-term debt financing with fixed duration and interest rates. Borrowing of debt capital is determined given the strategic decisions on financial structuring with respect to equity funding, dividend policy, and financial leverage (Walters, 1999). Cash flows from fixed asset and debt management are integrated into the cash flow balance of the model.

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

563

with single-stage production and constrained capacity. The following notation is used: Sets and indices pAP l A LOp ðp,lÞ A PLOp t, t s,u A S

products operations locations valid product–location combinations for operations locations periods scenarios

Fig. 3. Illustration of the postponement approach.

Capacity-related decisions and decisions on mid-term financial planning involve positive lead times as well as low reversibility and can have substantial impact on overall value creation. Structural decisions are therefore taken here-and-now based on imperfect information about future developments, and the corresponding plans should be stable irrespective of the scenario realized (Birge and Louveaux, 1997). In contrast, planning of sales and production/distribution quantities as well as shortterm liquidity can follow a wait-and-see approach with flexible contingency plans postponing control decisions until the scenario realized is revealed (Mulvey et al., 1995). Robust decision-making and two-stage stochastic programming provide corresponding frameworks to manage the two dichotomies stability vs. flexibility and here-and-now vs. wait-and-see (Gupta and Rosenhead, 1968; Birge and Louveaux, 1997; Scholl, 2001). However, the strict separation of decisions into stable first-stage and flexible second-stage variables for the entire planning period is a considerable restriction, especially in multi-period approaches precluding partial postponement and revision of structural decisions. In our approach, we suspend this strict assumption and require stability of structural first-stage decisions only for the shorter frozen period compared to the entire planning period (Gupta and Rosenhead, 1968). Structural decisions beyond the frozen horizon can be postponed to benefit from updated information in a rolling horizons approach (Benjaafar et al., 1995). This robust postponement approach is implemented for capacityrelated decisions and decisions on mid-term debt management. Scenario-specific decision variables are utilized for this purpose, but identical values for all scenarios are required for the frozen period. Remaining decisions can be postponed and are also modeled as flexible second-stage variables. Depending on the length of the frozen period, the robust compromise focuses more on stability or flexibility of the solution. With a frozen period of 0, the postponement approach turns into the totally flexible wait-and-see approach. A totally stable here-and-now approach evolves if the frozen horizon equals the planning horizon. Fig. 3 illustrates the postponement approach as a robust generalization of multi-period two-stage stochastic programming.

Decision variables stock of technical equipment at location l TElstt for scenario s at the end of period t acquired in period t IV lst ,DV lstt (Dis-)investment in technical equipment at location l for scenario s in period t (acquired in period t) Olst overtime at operations location l for scenario s in period t Iplst inventory of product p at operations location l for scenario s at the end of period t DLst amount of long-term debt for scenario s borrowed in period t Auxiliary variables Economic Value Added, negative deviation EVAs ,Ds of EVA for scenario s TCM st ,DPst total contribution margin, depreciation and capital loss of disinvestments for scenario s in period t CAst position in current net assets for scenario s at the end of period t capX lst normal capacity, overtime capacity at location l for scenario s in period t OCF st ,OM st , cash flow from operations, open items, FMst ,AM st ,DM st short-term financial investments, fixed assets, debt management for scenario s in period t Parameters pr s

g t 0 ,t 1 t h ,T DL tTE 0 , t0 z fa,fc ihr ,iDL ul,dd

4. Decision model In the following, we extend the decision model of Hahn and Kuhn (2011b) according to the conceptual approach developed in the previous section. The objective function follows the integrated approach to value-based performance and risk optimization as described in Section 3.2. We implement the decision variables and constraints for robust capacity adjustment and mid-term financial planning according to the concepts of Section 3.3. Discrete scenarios with respective probabilities are utilized to consider demand uncertainty in a make-to-stock supply chain

cap sv,cl

TE0lt TETl ,TEmax l ect DL0t

probability of scenario s risk preference parameter last period before the planning period, first period of the planning period frozen horizon, planning horizon earliest period of acquired technical equipment, borrowed long-term debt tax rate average balance of residual fixed assets, fixed costs hurdle rate, interest rate for long-term debt capital useful lifetime of technical equipment, duration of long-term debt capital capacity of one machine salvage value of technical equipment, capital loss factor of technical equipment sold initial stock of technical equipment at location l acquired in period t closing stock, maximum stock of technical equipment at location l exogenous cash flow in period t initial balance of long-term debt borrowed in period t

564

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

4.1. Objective function The objective function for integrated performance and risk optimization in (2) is linearized as described in (3) and (4). D denotes the negative deviation of EVA from the aspiration level 0 in scenario s. X X pr s  EVAs ð1gÞ  prs  Ds ð3Þ max g  sAS

sAS

EVAs þ Ds Z0 Ds Z0

8s A S

ð4Þ

8s A S

ð5Þ

According to the definition in (1) and the value driver tree in Fig. 1, EVA for scenario s is implemented in (6) as the difference between net operating profit after tax (NOPAT) and the capital charge. NOPAT results from total contribution margin TCM for scenario s in period t, depreciation DP for scenario s in period t, and fixed costs fc considering tax rate z. Depreciations become variable costs since (dis-)investments in technical equipment are decision variables. Total contribution margin TCM is calculated from net sales deducting sales costs and variable costs of operations, taking into account changes in inventory. TCM corresponds to net profits in the value driver tree and expenses include depreciation as well as fixed costs. 0 BX B T EVAs B ðTCM st DPst fcÞ  ð1zÞ B @ t ¼ t1 |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} 0

1 1

t1 X X

capital charge

8s A S

ð6Þ

The capital charge is derived from net assets and weighted average cost of capital (WACC) in the value driver tree (see Fig. 1). To avoid misinterpretations with respect to risk considerations within the WACC (Young and O’Byrne, 2001), we assume an externally predefined hurdle rate ihr to calculate EVA. Net assets include the average balance of residual fixed assets fa, the net value of technical equipment TE at location l for scenario s at the end of period t acquired in period t, and the position in current net assets CA for scenario s at the end of period t. The net value of technical equipment results from the net book value bv at the end of lifetime period t and the salvage value sv. Current net assets CA for scenario s at the end of period t cover short-term financial investments, inventories, accounts receivable, and cash. Accounts payable are deducted since they are non-interest-bearing debt capital (Kaplan and Atkinson, 1998). 0 t1 X X DPst @ TElstt  ðbvt1t bvtt Þ l A LOp t ¼ tTE 0

þ

t1 X X

TElst1, t DV lstt TElstt ¼ 0 8l A LOp ; 8s A S; t ¼ t 1 : :T; t ¼ tTE 0 : :ðt1Þ 8l A LOp ; 8s A S; t ¼ t 1 : :T

TElstt r TEmax l

8l A LOp ; 8s A S; t ¼ t 1 : :T

ð8Þ ð9Þ ð10Þ

t ¼ tTE 0

Eqs. (8) and (9) cover mass balances of technical equipment TE at location l for scenario s at the end of period t acquired before and in the current period t. IV at location l for scenario s denotes investments in period t. Since these investments cannot be immediately sold in the same period, investments do not need to be considered in (8). In (10), TEmax defines the maximum number of technical equipment TE at location l. t1 X

capX lst 

TElst1, t  cap ¼ 0

8l A LOp ; 8s A S; t ¼ t 1 : :T

ð11Þ

t ¼ tul

TElst0 t ¼ TE0lt T X

8l A LOp ; 8s A S; t ¼ tTE 0 : :t 0

TElsT t ¼ TETl

8l A LOp ; 8s A S

ð12Þ

ð13Þ

t ¼ Tul þ 1

Normal production capacity capX at location l for scenario s in period t is calculated in (11) from the number of available technical equipment at the end of the previous period t1 and the capacity cap of a machine. Initial and closing stock of technical equipment are defined in (12) and (13). IV lst IV lut ¼ 0

8l A LOp ; 8s,u A S; u a s; t ¼ t 1 : :t h

ð14Þ

DV lstt DV lutt ¼ 0

1

8l A LOp ; 8s,u A S; u a s; t ¼ t 1 : :t h ; t ¼ tTE 0 : :ðt1Þ

DV lstt  ðbvt1t bvtt  ð1clÞÞA ¼ 0

l A LOp t ¼ tTE 0

8s A S; t ¼ t 1 : :T

The model calculates sales quantities and the amount of marketing activities. Sales quantities are limited by actual customer demand, but can be extended within limits using marketing activities. Procurement, production, and transportation as well as storage quantities are also derived considering production and storage capacity. Production capacity can be extended using overtime provided by subcontractors. Sales prices, cost prices, and cost unit rates are fixed due to long-term contracts. Mass balance equations are required to calculate production quantities and raw materials according to the direct demand coefficients from the bill of materials. Initial and closing inventories of products and materials are considered for the rolling horizons approach. Lead times are implemented using predefined levels of inventory for final products and raw materials that have to already be in stock at the end of the previous period.

t X

C C hr C A TElst1t  ðbvt1t þ svÞ þ CAst1  i C ¼ 0  C t ¼ t1 A l A LOp t ¼ tTE 0 |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} @fa þ

4.2. Constraints for the physical domain

IV lst TElstt ¼ 0

net operating profit after tax ðNOPATÞ

T X

period t. The capital loss due to a lower market value compared to the net book value is considered utilizing the time-independent loss factor cl. Remaining formulae to complete the objective function can be found in Hahn and Kuhn (2011b).

ð7Þ

In (7), depreciation DP for scenario s in period t covers regular depreciation on the technical equipment TE at location l for scenario s at the end of period t acquired in period t. bv at the end of lifetime period t denotes the net book value excluding the salvage value. The second term comprises regular depreciation of disinvested technical equipment DV at location l for scenario s in period t acquired in

ð15Þ

In (14) and (15), (dis-)investments IV and DV at location l in period t (acquired in period t) are required to be identical for all scenarios s to obtain scenario-independent results up to the frozen horizon t h . Scenario-independent decisions for overtime O at location l in period t and inventories I for product p at location l at the end of period t within the frozen period are implemented analogously in (16) and (17). Olst Olut ¼ 0

8l A LOp ; 8s,u A S; u as; t ¼ t 1 : :t h

ð16Þ

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

Iplst Iplut ¼ 0

8ðp,lÞ A PLOp ; 8s,u A S; u a s; t ¼ t 1 : :t h

ð17Þ

(Dis-)investments IV and DV at location l for scenario s in period t (acquired in period t) are non-negative integer variables in (18). Thus the number of technical equipment TE at location l for scenario s at the end of period t acquired in period t can be defined in (19) as a non-negative real decision variable. Remaining formulae to complete the physical domain can be found in Hahn and Kuhn (2011b). IV lst ,DV lstt A N

8l A LOp ; 8s A S; t ¼ t 1 : :T; t ¼ tTE 0 : :T

8l A LOp ; 8s A S; t ¼ t 0 : :T; t ¼ tTE 0 : :T

TElstt Z0

ð18Þ ð19Þ

4.3. Constraints for the financial domain Financial investments and short-term borrowing are covered with a one-period horizon considering a given bank line of credit. Open items from accounts receivable and payable have a payment term of one period, but factoring and early payment deducting cash discount can be used to manage liquidity. Financial balances are initialized according to the balance sheet. Cash flows from fixed asset and debt management (AM, DM) are added to the mass balance equation of cash C in (20). OCF, OM, and FM denote cash flows from operations, open items, and short-term financial investments for scenario s in period t. ec in period t contains the exogenous cash flow due to obligations from fixed contracts, taxes, and dividends. C st1 OCF st þOMst þ FM st þ AM st þ DM st C st ¼ ect 8sA S; t ¼ t 1 : :T ð20Þ 0 AMst @

t1 X X

l A LOp t ¼ tTE 0



X

1 8s A S; t ¼ t 1 : :T

ð21Þ

l A LOp

In (21), cash flows from asset management AM for scenario s in period t result from (dis-)investments evaluated at the salvage value sv plus net book value bv considering the capital loss factor cl. Eq. (22) covers cash flows from debt management DM for scenario s in period t including paid interests at the interest rate iDL . The initial balance of long-term debts is defined in (23). ! t1 X DL DM st  DLst DLstdd  DLst  i ¼ 0 8s A S; t ¼ t 1 : :T t ¼ tdd

ð22Þ DLst ¼ DL0t

8s A S; t ¼ tDL 0 : :t 0

period of 13 periods (each equal to four weeks), and thus include one seasonal cycle and one financial year. Three consumer goods products (P1–P3) are manufactured in two plants (F1, F2) using a single-stage production process. Two suppliers (Z1, Z2) provide the raw materials (R1–R3) required. Final products are stored in two warehouses (W1, W2) before being delivered to five sales markets (M1–M5). Master data regarding the work plan and the bill of material, sales prices and unit costs, initial and target inventories as well as financial balances are adapted from the case-oriented example in Hahn and Kuhn (2011b). The supply chain network is depicted in Fig. 4. The capacity of 600,000 capacity units (cu) per plant results from six machines with a capacity of 100,000 cu each. Normal capacity can be extended by up to 20% using overtime. There are plans to purchase two machines per plant in the current financial year. The acquisition price of a machine with a useful lifetime of 39 periods (three financial years) is 3 million monetary units (mu). Relevant book values over the useful lifetime are derived according to straight-line depreciation considering a salvage value of 200,000 mu. The age of the current machines in each plant is 20 periods. The average balance of fixed assets excluding technical equipment amounts to 38 million mu; fixed costs per period are 3.5 million mu. Long-term debts of 10 million mu with a duration of 13 periods (one financial year) mature in periods 3 and 7, amounting to 5 million mu, respectively. The interest rate for long-term debt equals 0.5% per period. A moderate risk-averse decision-maker with g ¼ 0:3 is assumed. The frozen horizon is implemented in period 3 to ensure sufficient planning flexibility for periods 4–13. b

DV lstt  ðbvtt  ð1clÞ þ svÞ

IV lst  ðbv0 þsvÞA ¼ 0

ð23Þ

dplst ¼ sf s  dpl      2p T1 þtr  t þlet  tþ  1 þamp  cos 2 T

DLst Z 0

8s,u A S; u a s; t ¼ t 1 : :t h

8s A S; t ¼ tDL 0 : :T

ð24Þ ð25Þ

5. A case-oriented example 5.1. Base case The case-oriented example considers a consumer goods manufacturer with a centralized S&OP function. We cover a planning

ð26Þ

Demand data d for product p at sales location l for scenario s in period t is assumed to be correctly specified by the forecast model b in (26). A base level of customer demand d for product p at sales location l is adapted due to further demand components. Seasonality can be described using a harmonic oscillation with a seasonal amplitude amp and a peak in the middle of the seasonal cycle. The linear trend tr and the level shift le in period t account for gradual and one-time changes in market demand. We assume that the demand structure is constant over time, but different scenarios with corresponding probabilities can be observed, raising or lowering overall demand according to the scenario factor sf for scenario s. According to Hahn and Kuhn (2011b), we apply a robust scenario set of size 5 to keep contingency plans to a manageable

In (24), decisions on long-term borrowing are required to be identical for all scenarios to obtain scenario-independent results for the frozen period. Long-term debt DL for scenario s borrowed in period t is defined over the non-negative domain in (25). Remaining formulae to complete the financial domain can be found in Hahn and Kuhn (2011b). DLst DLut ¼ 0

565

Fig. 4. Supply chain layout and product allocation.

566

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

size and to sufficiently approximate the probability function of the scenario factors. The data for the base level of customer demand and the probability function (pr) of the scenario factors (sf) are provided in Tables 1 and 2, respectively. We assume a seasonal amplitude of 35%, linear demand growth of 2%, and no level shift over the planning period. For the numerical analysis, the decision model in Section 4 is implemented in IBM ILOG OPL v6.3. The model consists of 11,960 integer and 19,426 continuous decision variables as well as 30,415 constraints. Optimal solutions can be found using CPLEX v12.1 on a computer with a 2.13 GHz processor and 3 GB RAM within 37 s. Average EVA over the planning period amounts to 2.21 million mu (see Fig. 5), and is calculated from the NOPAT and the capital charge of 8.16 and 5.95 million mu, respectively. EVA values per period are within a range between  0.6 and þ0.7 million mu with a peak around period 7 reflecting the seasonal structure of demand (see Fig. 6). Preproduction in periods 1–4 raises inventories (see Fig. 7), and thus increases the capital charge. At the same time, NOPAT remains low since high levels of overtime induce additional costs and revenues are only realized when products are actually sold to the market. The average amount of overtime decreases in period 4 as can be seen in Fig. 6 since additional technical equipment is acquired for the positive scenarios.

Table 1 Base level of customer demand (units). Products

Sales locations

P1 P2 P3

The postponement approach for (dis-)investment planning derives contingency plans with investments for the intermediate scenario S3 and the positive scenarios S4 and S5. The number of technical equipment in each scenario over the planning period is depicted in Fig. 8. The purchase of four machines is predetermined by the long-term level. However, two additional machines are sufficient and further investments can thus be postponed to the planning horizon to avoid idle capacity. Due to the seasonal demand pattern, machines are even sold in scenarios S1 and S3 to reduce capacity costs. In general, the results confirm the importance of asset utilization as a major value driver and illustrate the interdependencies between capacity adjustments, overtime, and inventories. 5.2. Further numerical analyses

M1

M2

M3

M4

M5

32,000 48,000 60,000

27,000 49,000 65,000

30,000 52,000 57,000

32,000 52,000 60,000

34,000 55,000 64,000

Table 2 Discrete probability function (pr) of the scenario factors (sf). Scenarios

pr sf

Fig. 6. Numerical results: EVA per period.

S1

S2

S3

S4

S5

0.10 0.70

0.20 0.85

0.30 0.95

0.30 1.05

0.10 1.20

Fig. 5. Numerical results: value driver tree of EVA (million mu).

In the first analysis, we examine the economic value of flexible options in capacity (dis-)investment planning. The amount of available overtime and the size of incremental capacity additions are used for a sensitivity analysis. Increasing available overtime from 20 to 50% of normal capacity reduces lost sales by 1.1 pp from 8.1 to 7.0%. As a consequence, EVA increases by 32% on average from 2.21 to 2.92 million mu. Interestingly, increasing overtime beyond 40% of normal capacity does not show any effects. Comparable effects evolve in the event of small sizes per capacity addition. Reducing the size of technical equipment from 200,000 cu to 50,000 cu decreases lost sales by 2.4 pp from 10.4 to 8.0%, and enhances average EVA by 26% from 1.86 million to 2.34 million mu. Generalizing the results, flexible capacity options facilitate the timing of discrete capacity (dis-)investments when demand grows gradually, and thus create additional economic value. The results are summarized in Tables 3 and 4. Secondly, we extend the previous investigations and compare the results of the integrated approach to a sequential approach where decisions on capacity adjustment timing are determined before mid-term S&OP is conducted. For this purpose, we predetermine the capacity adjustment decision of a specific period and optimize the remaining decision variables. This approach is iterated for each period up to the planning horizon and an average result for the entire planning period is derived. Besides the linear demand growth of 2% per period, we examine a second demand pattern: a level shift in demand of 15,000 units per product and sales location occurring in period 5 when seasonal demand is also rising. Both demand patterns result in comparable demand increases for the entire planning period. The improvement potential due to integrated planning is between 18.5 and 27.9% (level shift) as well as 35.1 and 69.4% (linear trend) on average in terms of EVA. In the event of a linear trend, the gap is higher since demand changes occur gradually over time compared to a one-time level shift, and thus provide

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

567

Fig. 7. Numerical results: overtime and inventories per period.

Table 5 Improvement potential of the integrated approach (available overtime). Overtime (% of normal capacity)

20 30 40 50

Improvement potential (%) Level shift

Linear trend

25.1 20.3 20.2 18.5

54.6 44.4 40.0 35.1

Table 6 Improvement potential of the integrated approach (incremental capacity). Fig. 8. Numerical results: number of technical equipment per period. Incremental capacity (’000 cu)

Table 3 Results of increasing available overtime. Overtime (% of normal capacity)

Avg. lost sales (%)

Avg. EVA (million mu)

20 30 40 50

8.1 7.7 7.0 7.0

2.21 2.69 2.92 2.92

200 100 66 50

Improvement potential (%) Level shift

Linear trend

27.1 25.1 27.9 25.0

69.4 54.6 52.5 50.0

Table 4 Results of decreasing size of incremental capacity. Incremental capacity (’000 cu)

Avg. lost sales (%)

Avg. EVA (million mu)

200 100 66 50

10.4 8.1 7.7 8.0

1.86 2.21 2.45 2.34

more potential for capacity adjustment timing in the integrated approach. When capacity flexibility is high (available overtime 50%, incremental capacity 50,000 cu), the impact of integrated adjustment timing is lower. The improvement potential due to the integrated approach decreases between 2.1 and 6.6 pp (level shift) as well as 19.4 and 19.5 pp (linear trend) on average. The results summarized in Tables 5 and 6 underline the importance of integrated capacity adjustment timing and S&OP to optimize asset utilization as a major mid-term value driver. In the third analysis, we evaluate the postponement (PP) approach to illustrate the impact of additional planning flexibility. We use the base case as introduced above and combine two different demand patterns: a level shift in demand of 7500 units per product and sales location occurring in period 5 as well as linear demand growth of 1% per period. We investigate three different frozen horizons (0, 3, and 13) and the range of weakly

Fig. 9. Results for the postponement approach.

risk-averse preferences for g A ð0; 1. Upside potential (UP) and downside risk (DR) are derived from the distribution of EVA calculating UP as the complementary upper partial moment to DR (see Section 3.2). UP and DR values are depicted in Fig. 9 for different risk preferences and frozen horizons applying the PP approach as introduced above. The wait-and-see (WS) and here-and-now (HN) approaches evolve as limiting cases of the PP approach for frozen horizons of 0 and 13. For a specific individual risk preference, the gap

568

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

between the WS and HN approach added up for UP and DR equals the expected value of perfect information (EVPI). The EVPI measures the impact of uncertainty and corresponds to the maximum amount a decision-maker is willing to pay for perfect information (Birge and Louveaux, 1997). Bringing forward the frozen horizon from period 13 (HN approach) to 3 (PP approach) reduces the impact of uncertainty since upside potential increases by 1.6 million mu on average and downside risk marginally decreases by 80,000 mu on average. Since the robust framework immunizes solutions against the impact of negative scenarios, the flexibility introduced in the postponement approach therefore mainly results in additional upside potential. Investigating the detailed (dis-)investment decisions confirms this hypothesis: capacity adjustments are aligned with the negative scenarios to avoid costly idle capacity. Postponed decisions can be adapted accordingly in subsequent planning iterations in the event of a positive scenario. In general, the postponement approach improves expected value creation in terms of EVA and restricts the impact of uncertainty since it reduces the gap vs. the WS approach. The decision-maker balances upside potential and downside risk to determine individual risk preference and the corresponding optimal plan using Fig. 9. A highly risk-averse decision-maker focuses on objective robustness and minimizes downside risk realizing a substantial loss in upside potential. By contrast, a riskneutral decision-maker aims for solution robustness and an optimal trade-off between upside potential and downside risk.

6. Conclusion and outlook A decision framework for simultaneous investment, operations, and financial planning in supply chains is presented using an integrated approach for value-based performance and risk management. The decision model implements the Economic Value Added (EVA) concept as a prevalent metric of value-based performance. The approach supports integrated decision-making for mid-term S&OP and capacity adjustment timing based on long-term instructions from strategic network planning. A unified approach for capacity adjustment and equipment replacement planning is developed. Robust optimization methods are applied to mitigate operational risk impact and implement a postponement approach for capacity planning. The benefits of the approach are highlighted using a case-oriented example. Practical implications from the case-oriented example can be summarized as follows: asset utilization is confirmed as a major value driver at the mid-term level, and a comprehensive approach is thus required to align different levers of capacity supply and demand management. Integrated planning of capacity adjustment timing and capacity control in S&OP reveals considerable improvement potential compared to a conventional sequential approach. Capacity adjustment timing is critical in the event of limited flexible capacity options as well as gradually changing aggregate demand. Postponing (dis-)investment decisions in a rolling horizons approach increases flexibility and creates additional upside potential in terms of EVA since the robust optimization approach hedges against the impact of negative scenarios. Although selective matters are simplified in the case-oriented example to focus the numerical analyses on the aspects of our research, the underlying assumptions and fundamental relationships still remain valid. Taking a more general perspective on the business problem in focus reveals some limitations of the approach presented. At least three parameters of the decision model could be subject to management considerations that would change the basic conditions of the problem: the maximum level of overtime capacity,

the lead time for (dis-)investments, and the level of accuracy of the demand information. Increasing flexible overtime capacity could improve responsiveness towards customers resulting in a higher operating profit margin. Covering different manufacturing technologies or collaborating with a leasing company could reduce lead times for adjustments in technical equipment, allowing for a better match of capacity supply and demand. Although the expected value of perfect information (EVPI) is calculated in the approach to quantify the impact of incomplete information, the decision model does not consider the benefits and costs of further information gathering. Even though the decision model and the case-oriented example originate from the consumer goods industry, the approach could be also applied to companies in other industries, such as chemicals, pulp and paper, and metals, with a centralized planning process/function responsible for S&OP. Specialized SCM software vendors provide corresponding standard software for integrated business planning using mixed-integer linear programming to optimize financial performance metrics in sales, operations, and working capital management. However, the software packages do not apply robust optimization methods to provide decision support for risk management and do not cover capacity adjustment and equipment replacement planning. From our point of view, practitioners would strongly benefit from the corresponding functionality being available as standard packaged software. Integrated capacity management improves asset utilization and reduces lost sales, ultimately creating additional economic value. Real decision support for risk management would be a substantial improvement since simple scenario analysis does not provide a robust and implementable solution across all scenarios. The impact of further parameters should be evaluated with respect to future research. The capital loss factor and different depreciation methods can have considerable impact on (dis-)investment decisions. Furthermore, the approach presented in this paper can be extended in several directions. Considerations regarding technology selection and economies of scale could lead to interesting results regarding specificity and size of investments as well as hedging strategies in capacity (dis-)investment planning. Extending the decision model towards a hierarchical planning framework by introducing a short-term planning level below mid-term S&OP would be a more rigorous approach. This would allow investigation of the effects of detailed lot-sizing in production and distribution planning as well as the implications of short-term financial planning in supply chains.

Acknowledgments We thank the three anonymous referees for their valuable feedback which has helped to improve the paper substantially.

References Aghezzaf, E., 2005. Capacity planning and warehouse location in supply chains with uncertain demands. Journal of the Operational Research Society 56 (4), 453–462. Ahmed, S., King, A.J., Parija, G., 2003. A multi-stage stochastic integer programming approach for capacity expansion under uncertainty. Journal of Global Optimization 26 (1), 3–24. Barahona, F., Bermon, S., Gunluk, O., Hood, S., 2005. Robust capacity planning in semiconductor manufacturing. Naval Research Logistics 52 (5), 459–468. Barbaro, A., Bagajewicz, M.J., 2004. Managing financial risk in planning under uncertainty. AIChE Journal 50 (5), 963–989. Benjaafar, S., Morin, T.L., Talavage, J.J., 1995. The strategic value of flexibility in sequential decision making. European Journal of Operational Research 82 (3), 438–457.

G.J. Hahn, H. Kuhn / Int. J. Production Economics 140 (2012) 559–569

Bhutta, K.S., Huq, F., Frazier, G., Mohamed, Z., 2003. An integrated location, production, distribution and investment model for a multinational corporation. International Journal of Production Economics 86 (3), 201–216. Birge, J.R., Louveaux, F., 1997. Introduction to Stochastic Programming. Springer, New York. Bok, J.-K., Lee, H., Park, S., 1998. Robust investment model for long-range capacity expansion of chemical processing networks under uncertain demand forecast scenarios. Computers & Chemical Engineering 22 (7–8), 1037–1049. Bradley, J.R., Arntzen, B.C., 1999. The simultaneous planning of production, capacity, and inventory in seasonal demand environments. Operations Research 47 (6), 795–806. Buxey, G., 2003. Strategy not tactics drives aggregate planning: structuring and planning operations. International Journal of Production Economics 85 (3), 331–346. Cai, J., Liu, X., Xiao, Z., Liu, J., 2009. Improving supply chain performance management: a systematic approach to analyzing iterative KPI accomplishment. Decision Support Systems 46 (2), 512–521. Chen, Z.-L., Li, S., Tirupati, D., 2002. A scenario-based stochastic programming approach for technology and capacity planning. Computers & Operations Research 29 (7), 781–806. Christopher, M., Ryals, L., 1999. Supply chain strategy: its impact on shareholder value. International Journal of Logistics Management 10 (1), 1–10. Eppen, G.D., Martin, R.K., Schrage, L., 1989. A scenario approach to capacity planning. Operations Research 37 (4), 517–527. Fleischmann, B., Meyr, H., Wagner, M., 2008. Advanced planning. In: Stadtler, H., Kilger, C. (Eds.), Supply Chain Management and Advanced Planning. Springer, Berlin, pp. 81–106. Goetschalckx, M., Fleischmann, B., 2008. Strategic network design. In: Stadtler, H., Kilger, C. (Eds.), Supply Chain Management and Advanced Planning. Springer, Berlin, pp. 117–132. Gunasekaran, A., Patel, C., McGaughey, R.E., 2004. A framework for supply chain performance measurement. International Journal of Production Economics 87 (87), 333–347. Gupta, S.K., Rosenhead, J., 1968. Robustness in sequential investment decisions. Management Science 15 (2), B18–B29. Hahn, G.J., Kuhn, H., 2011a. Optimising a value-based performance indicator in mid-term sales and operations planning. Journal of the Operational Research Society 62 (3), 515–525. Hahn, G.J., Kuhn, H., 2011b. Value-based performance and risk management: a robust optimization approach. International Journal of Production Economics, doi:10.1016/j.ijpe.2011.04.002. Hsu, C.-I., Li, H.-C., 2009. An integrated plant capacity and production planning model for high-tech manufacturing firms with economies of scale. International Journal of Production Economics 118 (2), 486–500. Huang, K., Ahmed, S., 2009. The value of multistage stochastic programming in capacity planning under uncertainty. Operations Research 57 (4), 893–904. Julka, N., Baines, T., Tjahjono, B., Lendermann, P., Vitanov, V., 2007. A review of multi-factor capacity expansion models for manufacturing plants: searching for a holistic decision aid. International Journal of Production Economics 106 (2), 607–621. Kaplan, R.S., Atkinson, A.A., 1998. Advanced Management Accounting, third ed. Prentice-Hall, Upper Saddle River. Kleijnen, J.P.C., Smits, M.T., 2004. Performance metrics in supply chain management. Journal of the Operational Research Society 54 (5), 507–514. Lainez, J.M., Puigjaner, L., Reklaitis, G.V., 2009. Financial and financial engineering considerations in supply chain and product development pipeline management. Computers & Chemical Engineering 33 (12), 1999–2011. Lambert, D.M., Pohlen, T.L., 2001. Supply chain metrics. International Journal of Logistics Management 12 (1), 1–20. Lavaja, J., Adler, A., Jones, J., Pham, T., Smart, K., Splinter, D., Steele, M., Bagajewicz, M.J., 2006. Financial risk management for investment planning of new commodities considering plant location and budgeting. Industrial & Engineering Chemistry Research 45 (22), 7582–7591.

569

Li, S., Tirupati, D., 1994. Dynamic capacity expansion problem with multiple products: technology selection and timing of capacity additions. Operations Research 42 (5), 958–976. Liberatore, M.J., Miller, T., 1998. A framework for integrating activity-based costing and the balanced scorecard into the logistics strategy development and monitoring process. Journal of Business Logistics 19 (2), 131–154. Luss, H., 1982. Operations research and capacity expansion problems: a survey. Operations Research 30 (5), 907–947. Majumdar, S., Chattopadhyay, D., 1999. A model for integrated analysis of generation capacity expansion and financial planning. IEEE Transactions on Power Systems 14 (2), 466–471. Mulvey, J.M., Vanderbei, R.J., Zenios, S.J., 1995. Robust optimization of large-scale systems. Operations Research 43 (2), 264–281. Nawrocki, D., 1999. A brief history of downside risk measures. Journal of Investing 8 (3), 9–25. Nobes, C., Parker, R.B., 2008. Comparative International Accounting, tenth ed. Prentice-Hall, Harlow. Ojala, M., Hallikas, J., 2006. Investment decision-making in supplier networks: management of risk. International Journal of Production Economics 104 (1), 201–213. Olhager, J., Rudberg, M., Wikner, J., 2001. Long-term capacity management: linking the perspectives from manufacturing strategy and sales and operations planning. International Journal of Production Economics 69 (2), 215–225. Paraskevopoulos, D., Karakitsos, E., Rustem, B., 1991. Robust capacity planning under uncertainty. Management Science 37 (7), 787–800. ¨ T., Sandstrom, ¨ Pirttila, J., 1995. Manufacturing strategy and capital budgeting process. International Journal of Production Economics 41 (1–3), 335–341. Poojari, C.A., Lucas, C., Mitra, G., 2008. Robust solutions and risk measures for a supply chain planning problem under uncertainty. Journal of the Operational Research Society 59 (1), 2–12. Rajagopalan, S., 1998. Capacity expansion and equipment replacement: a unified approach. Operations Research 46 (6), 846–857. Rajagopalan, S., Swaminathan, J.M., 2001. A coordinated production planning model with capacity expansion and inventory management. Management Science 47 (11), 1562–1580. Rappaport, A., 1998. Creating Shareholder Value: A Guide for Managers and Investors, second ed. Free Press, New York. ¨ Rastogi, A.P., Fowler, J.W., Carlyle, W.M., Araz, O.M., Maltz, A., Buke, B., 2011. Supply network capacity planning for semiconductor manufacturing with uncertain demand and correlation in demand considerations. International Journal of Production Economics 134 (2), 322–332. Ritchie, B., Brindley, C., 2007. Supply chain risk management and performance. International Journal of Physical Distribution & Logistics Management 27 (3), 303–322. Scholl, A., 2001. Robuste Planung und Optimierung: Grundlagen, Konzepte und Methoden, Experimentelle Untersuchungen. Physica, Heidelberg. Shapiro, J.F., 2007. Modeling the Supply Chain, second ed. Thomson-Brooks/Cole, Belmont. Supply Chain Council, 2010. Supply chain operations reference (SCOR) model: overview version 10.0 /http://supply-chain.org/f/SCOR-Overview-Web.pdfS. Tang, C.S., 2006. Perspectives in supply chain risk management. International Journal of Production Economics 103 (2), 451–488. van Mieghem, J.A., 2003. Capacity management, investment, and hedging: review and recent developments. Manufacturing & Service Operations Management 5 (4), 269–302. Vaughan, E.J., 1997. Risk Management. John Wiley, New York. Walters, D., 1999. The implications of shareholder value planning and management for logistics decision making. International Journal of Physical Distribution & Logistics Management 29 (4), 240–258. Wu, S.D., Erkoc, M., Karabuk, S., 2005. Managing capacity in the high-tech industry: a review of literature. The Engineering Economist 50 (2), 125–158. Young, S.D., O’Byrne, S.F., 2001. EVA and Value-based Management: A Practical Guide to Implementation. McGraw-Hill, New York.