Content uploaded by Mualla Gonca Avci
Author content
All content in this area was uploaded by Mualla Gonca Avci on Apr 14, 2023
Content may be subject to copyright.
e-ISSN 2757-5195
Çanakkale Onsekiz Mart University
Journal of Advanced Research in Natural and Applied Sciences
Open Access
doi.org/10.28979/jarnas.1105502
2022, Vol. 8, Issue 4, Pages: 624-640
dergipark.org.tr/tr/pub/jarnas
An Integrated Risk Management Framework for Global Supply Chains
Mualla Gonca Avci1,*
1 Department of Industrial Engineering, Engineering Faculty, Dokuz Eylul University, Izmir, Türkiye
Article History
Abstract − In this study, a risk management framework is developed to support risk management decisions in global
supply chains. The proposed framework covers all phases of risk management, namely, risk identification, risk miti-
gation and control. In the risk identification phase of the framework, the supply chain is decomposed into either
material-level or product-level sub-networks according to the decision maker’s preference. Afterwards, the most crit-
ical sub-network is modelled to evaluate different risk mitigation strategies. In particular, a combination of redun-
dancy and flexibility strategies is considered for risk mitigation. These strategies are evaluated by simulation models
in terms of their effectiveness and efficiency. While inventory holding cost is used as efficiency measure, effective-
ness of the strategies is measured by premium freight ratio. The proposed framework provides a comprehensive and
reliable decision support since it covers all phases of risk management and relies on quantitative data, and statistical
analysis in risk modelling. Moreover, it is flexible as it can be easily adapted to any change in supply chain environ-
ment and strategy. In order to show the applicability of the framework, a practical demonstration is presented for a
European automotive company. The results indicate that the proposed framework improves the supply chain perfor-
mance in terms of efficiency and effectiveness.
Received:
18.04.2022
Accepted:
26.06.2022
Published:
15.12.2022
Research Article
Keywords − Premium freight, simulation, supply chain risk management, TOPSIS,
1. Introduction
As a result of the advances in communication and transportation technologies, today’s supply chains have
geographically expanded. This fact drives inter-firm competition into global scale. Consequently, supply
chains adopt lean principles, and partners in supply chain become more connected to gain cost advantage.
However, these strategies make supply chains vulnerable in terms of supply chain risks. Today, an adverse
event affected a supply chain partner can influence the entire chain. Therefore, supply chain risk management
(SCRM) has gained importance in both industry and academia. Nowadays, firms must evaluate interdepend-
encies in their supply chain, identify their risks and measure likelihoods, effects and severities of the identified
risks. Firms should develop risk management plans to avoid, mitigate or control the identified risks (Tummala
and Schoenherr 2011).
In global supply chains, SCRM is challenging due to the complex supply chain structures and interrelation-
ships. In fact, global supply chains can be investigated by decomposing them into sub-systems under several
scales by using different viewpoints such as a geographical zone, a hazard level, a supplier group or a particular
type of product. Then, risk management activities can proceed conveniently for the pre-specified critical sub-
systems. In addition, risk management in global supply chains should be flexible so that it can be adapted
easily to rapid changes in supply chain environment and competition strategy. Furthermore, SCRM should
mainly rely on quantitative data to perform more realistic and reliable risk analysis.
1
gonca.yunusoglu@deu.edu.tr
*Corresponding Author
Null Line
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
625
This study proposes an integrated SCRM framework for operational supply chain risk management in global
scale. The proposed framework covers all phases of risk management, namely, risk identification, risk mitiga-
tion, risk monitoring and control phases. Particularly, the framework involves a risk identification phase in
which the supply chain is decomposed into material-level or product-level sub-networks according to the de-
cision maker’s preference. Afterwards, risk mitigation strategies are simulated for the specified critical sub-
network, and the best factor levels for risk mitigation strategies are determined by an experimental design
approach. In particular, redundancy and flexibility strategies are considered for risk mitigation. In this study,
holding safety stock and excess production which is more than forecasted demand size are considered as re-
dundancy strategies. Additionally, supplier’s quantity flexibility is the flexibility strategy. These strategies are
considered in combined manner and evaluated in terms of both efficiency and effectiveness. Herein, the annual
holding cost of the supply chain is employed as a measure of efficiency. In addition, the ratio of premium
freights to regular orders is used to measure the effectiveness of the strategies. Specifically, in case of a short-
age or delay risk, requesting a premium freight is an effective contingency strategy. However, premium freight
is a type of last-minute shipment transported by airlines. It incurs very high costs to the responsible agent in a
short time frame due to its setup cost and transportation mode. Additionally, it is an indicator of vulnerability
of the supply chain. Hence, premium freight is used as an additional performance measure in this study to
measure the effectiveness of the strategies.
The proposed framework consisting of all phases of SCRM provides a comprehensive decision support for
SCRM unlike the majority of the studies in this field. The proposed framework is a reliable and realistic tool
as it uses quantitative data and employs statistical risk models relying on real historical data. Additionally, it
is flexible in determining the focus of risk management and can be easily adapted to any changes in supply
chain management strategy or environment. Furthermore, to our knowledge, premium freight is used as an
effectiveness measure for the first time for risk mitigation strategies in SCRM field.
The rest of the paper is organized as in the following. In Section 2, related studies in literature are overviewed.
In Section 3, the proposed framework is presented. Section 4 presents an application of the proposed frame-
work to an automotive supply chain. In Section 5, results of the application are discussed, and managerial
implications are provided. Section 6 concludes the study.
2. The Related Literature
As SCRM is still an emerging research field, definition of risk concepts and risk mitigating strategies are
still unclear. Therefore, several review studies such as (Singhal, Agarwal, and Mittal 2011), (Tang and
Nurmaya Musa 2011), (Colicchia and Strozzi 2012), (Sodhi, Son, and Tang 2012), (Rangel, de Oliveira, and
Leite 2014), (Heckmann, Comes, and Nickel 2015), (Ho et al. 2015) and Pournader et al. (2020) have been
published with the aim of classifying the studies on SCRM. Additionally, the reader can find broad descriptions
of risk concepts and risk management strategies in (Chopra and Sodhi 2004), (Christopher and Peck 2004),
and (Sheffi 2005). In this section, recent quantitative studies related to SCRM field are presented.
The majority of the recent studies in the related field deal with the risk assessment phase of SCRM. In recent
studies, multi-criteria decision making techniques, mathematical programming approaches, system analysis
and simulation are utilized to assess supply chain risks. Among these approaches, multi-criteria decision
making techniques are the most widely used tools in risk assessment. Wang et al. (2012) utilize a fuzzy
analytical hierarchy process (AHP) model for risk assessment of implementing green initiatives in a fashion
supply chain. Chaudhuri, Mohanty, and Singh (2013) utilize Failure Mode and Effects Analysis (FMEA) to
prioritize the failure modes of vulnerable suppliers in new product development process. Chen and Wu (2013)
propose an FMEA to categorize existing suppliers and select new suppliers by considering risk. Samvedi, Jain,
and Chan (2013) utilize fuzzy AHP and fuzzy TOPSIS methods to obtain a supply chain risk index. Aqlan and
Lam (2015a) propose a risk assessment framework which employs Bow-Tie analysis and fuzzy inference
system for supply chains. In another study, Aqlan and Lam (2015b) quantify supply chain risks by Bow-Tie
analysis, and select mitigation strategy by an optimization model. Govindan and Jepsen (2015) model supply
chain uncertainties as intuitionistic fuzzy numbers and assess them via ELECTRE-C. Oliveira et al. (2022)
identify and assess the supply chain risks of a home-care service provider via FMEA.
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
626
There exist a few mathematical programming applications in risk assessment context. Cardoso et al. (2015)
develop a mathematical model for supply chain design and planning to assess resilience of alternative supply
chain structures under different disruption types. Klibi and Martel (2012) propose a risk modelling approach
considering random, hazardous and deeply uncertain events causing supply chain disruptions. They utilize a
Monte-Carlo approach to assess the disruption impact based on the descriptive models.
Recently, system analysis and simulation approaches have become popular in risk assessment. Ghadge et al.
(2013) develop a risk management framework by using systems approach to capture the dynamic
characteristics of supply chain risks. Bueno-Solano and Cedillo-Campos (2014) investigate the impact of
disruptions originated from terrorist acts via a system dynamics model. Wagner, Mizgier, and Arnez (2014)
propose Monte-Carlo approach to evaluate possible losses due to disruptions in the US offshore oil industry.
Guertler and Spinler (2015) demonstrate the criticality of the operational risks by using a system dynamics
model that assesses the intra-organizational dynamics of risks.
In view of the related body of knowledge, it can be stated that the number of studies building comprehensive
SCRM frameworks considering all phases of SCRM is limited. Among them, Giannakis and Louis (2011)
propose a multi-agent based decision support system for supply chain disruption management. Schmitt and
Singh (2012) analyse inventory placement and back-up strategies against supply chain risks by a simulation
model. Carvalho et al. (2012) use a supply chain simulation model to evaluate the effects of mitigation
strategies on performance of each supply chain entity under a set of scenarios. Rajesh and Ravi (2015) employ
a grey-DEMATEL approach to investigate cause-effect relationships between supply chain risk mitigation
strategies. Simchi-Levi et al. (2015) develop a SCRM framework for Ford Motor Company to identify new
risks, evaluate proactive risk mitigation plans, and derive optimal contingency plans. Oliveira et al. (2019)
propose hybrid and flexible simulation-based optimization models for SCRM. However, they do not present a
real world application of their model. Kara et al. (2020) present an integrated SCRM framework which
employs data mining algorithms. Talukder et al. (2021) propose a multi-indicator supply chain management
framework that provides leanness, agility, sustainability, and resilience in the dairy business.
In this study, supply chain risks related to the physical flow in global supply chain networks are considered.
In related literature, there exist a few study considering such large supply chain networks ((Chaudhuri,
Mohanty, and Singh 2013), (Klibi and Martel 2012), (Wagner, Mizgier, and Arnez 2014), (Simchi-Levi et al.
2015)). As global supply chains consist of several facilities spread on several countries, risk management in
these supply chains is challenging. A possible solution to overcome this challenge may be decomposition of
the supply chain network into sub-networks by a risk assessment procedure as Chaudhuri, Mohanty, and Singh
(2013) and Klibi and Martel (2012) do. Hence, in this study, the supply chain network is decomposed into
critical sub-networks to according to their risk level. Unlike the previous studies, the proposed framework has
a more comprehensive risk identification phase in which the supply chain can be decomposed into the single-
product level or single-material level sub-networks. Therefore, the risks can be assessed at material or product-
level by considering the preference of supply chain managers. Consequently, the proposed framework enables
the flexibility essential for the real-world SCRM practices.
As stated previously, the proposed framework consists of a risk mitigation phase in addition to risk
identification phase. In risk mitigation phase, redundancy and flexibility strategies are investigated to the aim
of effective and efficient SCRM. In particular, holding safety stock and excess production which is more than
forecasted demand size are considered as redundancy strategies. Additionally, supplier’s quantity flexibility is
considered as the flexibility strategy. These strategies have been utilized in previous studies. However, they
have not been evaluated from effectiveness and efficiency perspectives simultaneously. Effectiveness is the
ability to achieve a predefined goal in case of adverse conditions. Efficiency is to ensure minimal spending of
resources to reach the goal (Heckmann, Comes, and Nickel 2015). In this study, premium freight ratio and
annual holding cost are utilized as effectiveness and efficiency measures, respectively. To the best of our
knowledge, there exists no study in related literature considering premium freight as a SCRM performance
measure.
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
627
3. The Proposed Supply Chain Risk Management Framework
In this study, an integrated SCRM framework is developed to support SCRM decisions by considering
holding costs and premium freights. The proposed framework considers not only premium freights but also
supply chain costs to ensure both effectiveness and efficiency objectives in SCRM. Moreover, the proposed
framework covers all commonly acknowledged phases of SCRM namely, risk identification, risk mitigation
and risk monitoring and control. Furthermore, the proposed framework has the flexibility in executing material
or product-level risk analyses according to preference of the decision maker. The process flow diagram sum-
marizing the phases of the proposed framework is illustrated in Figure 1. In subsequent sections, general pro-
cess of the framework is explained through risk management phases.
3.1. Risk Identification
Firstly, the risks affecting supply and delivery processes of the supply chain are identified. Basically, supply
chain risks can be classified into inbound and outbound risks by considering physical flow direction. The
inbound risks are related to the supply-side adverse events such as supplier delivery delay, delivery quantity
loss and supplier disruptions. The outbound risks are arisen from customer-side adverse events such as
customer demand variability, product delivery delay and shifting customer demand. As stated previously, the
proposed framework has the flexibility in focusing on the inbound or outbound risks. In this context, inbound
and/or outbound risks are quantified according to the preference of the manager. In particular, the supply chain
risks are quantified through statistical models developed by using past order, premium freight and customer
sales data. The risk models are obtained by fitting the historical data to probability distributions.
Once the supply chain risks related to each agent are quantified, the most critical sub-network in terms of risk
is identified. The critical sub-network may be related to a material or a product type. To identify the critical
sub-network, the most critical material or product is determined. In this context, the criteria for critical sub-
network identification should be specified. In criteria determination, the adverse effects of risks related to the
focal material or product and criticality of them should be considered. Hence, the critical sub-network
identification stage involves multi-criteria decision making. Therefore, utilization of a multi-criteria decision
making technique in this stage is appropriate. TOPSIS (Hwang & Yoon, 1981) is a suitable technique for this
stage as it ranks the alternatives according to their criticality.
Assume that a multi-criteria decision making problem has m alternatives, and n criteria,
. The ratings associated with the alternatives with respect to each criterion is included in a decision
matrix, . Then, the ratings are normalized to form the normalized decision matrix.
In particular, in identification of material or product-level sub-networks, more than one decision matrices may
come into consideration. As the suppliers and customers may be connected with more than one plant in the
supply chain, more than one assessment for a criterion may come out. In this case, group decision making
approaches can be used to merge the decision matrices into single decision matrix.
The criteria weights are determined by using entropy weighting method (Deng, Yeh, & Willis, 2000). Entropy
weighting method considers intrinsic information in each criterion, and does not require decision maker's
judgment. In this sense, entropy weighting method is utilized to determine criteria weights to reduce human
effort in decision making.
The entropy value indicating the amount of information contained in each criterion is calculated as follows.
(3.1)
The degree of divergence () of the average intrinsic information related to each criterion is calculated as
follows.
(3.2)
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
628
The Focus of Risk Manager
Material-Level Risk Management Product-Level Risk Management
Risk Monitoring... Risk Identification
Risk Mitigation
Supplier Delivery
Performance
Database
Identify inbound risks Identify inbound and
outbound risks
Statistical
risk models
Quantify the risks Quantify the risks
Customer Sales
Database
Statistical
risk models
Critical material
identification (TOPSIS)
Criteria for
material-level
risk identification
Premium Freight
Database
Critical product
identification (TOPSIS)
Criteria for
product-level risk
identification
The critical
sub-network The critical
sub-network
Full factorial experimental
design
Simulation experiments
Safety stock and
supplier
flexibility levels
MINITAB Response
Optimizer
Annual holding cost
and premium freight
ratio values
The best safety
stock and supplier
flexibility levels
Apply the plan
Monitor supply chain performance
and review SCRM process
Full factorial experimental
design
Simulation experiments
Demand for. adj. factor,
safety stock and supplier
flexibility levels
MINITAB Response
Optimizer
Annual holding cost
and premium freight
ratio values
Apply the plan
Monitor supply chain performance
and review SCRM process
The best demand for. adj.
factor, safety stock and
supplier flexibility levels
Figure 1. Process flow diagram of the proposed framework
As the degree of divergence represents divergence of the ratings in terms of criterion j, higher degree of
divergence leads to higher criterion weight. In this sense, criteria weights are calculated as follows.
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
629
(3.3)
Aggregation procedure of TOPSIS is based on weighted Euclidian distances to negative and positive ideal
solution.
(3.4)
(3.5)
where is the distance of the alternative i to positive ideal solution, is the distance of alternative i to
negative ideal solution, is the positive ideal value for criterion j, and is the negative ideal value for
criterion j.
The overall criticality index for alternative i is calculated as follows.
(3.6)
In the proposed framework, the overall criticality index obtained by TOPSIS present the criticality of material
or product in terms of risk. The material (or product) with the highest level of overall criticality is identified
as the most critical material (or product). The network related to the most critical material or product is
specified as the critical sub-network. Afterwards, the critical sub-network is modelled to develop and evaluate
risk mitigation strategies. Concept of the critical sub-network identification is illustrated with an example in
Figure 2. An example supply chain system is demonstrated in Figure 2. In case of material-level analysis, the
most critical material is identified as the material consumed by P1, P2, P3, and supplied by S2. Hence, the sub-
network consisting of S2, P1, P2, and P3 is identified as the critical sub-network. Alternatively, in case of
product-level analysis, the most critical product is identified as the product demanded by C3. Therefore, the
critical sub-network consists of the customer demanding the product (C3), the plant producing the product (P3)
and the suppliers supplying the materials required for production of the product (S2 and S4).
3.2 Risk Mitigation
In this stage, a discrete event system simulation model of the critical sub-network is developed. By utilizing
the simulation model, a number of risk mitigation strategies are evaluated in terms of their effectiveness and
efficiency. In evaluation of the risk mitigation strategies, interrelationships between the strategies must be
taken into account. For example, there is a strong relationship between redundancy and flexibility strategies.
These strategies are effective in mitigating supply chain risks. However, combinations of these strategies often
yield more effective and efficient risk mitigation due to their systemic effects. Herein, we cannot conclude that
one strategy is superior to another strategy in terms of both effectiveness and efficiency. Therefore, in this
study, these strategies are quantitatively described and evaluated by simulation experiments. The best
combination of these strategies is determined in terms of effectiveness and efficiency. To evaluate the
effectiveness and efficiency together, a multi-objective evaluation is required. Hence, both cost and premium
freight performance are considered to measure efficiency and effectiveness together. To observe the effects of
risk mitigation parameter levels a full factorial experiment is designed. Minitab Response Optimizer Tool is
used to obtain the best parameter levels that give minimum annual holding cost and premium freight ratio.
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
630
Figure 2. An example to critical sub-network identification
3.3 Risk Monitoring and Control
After the risk mitigation strategies are determined, outcomes of these strategies are continually monitored
and reviewed. To ensure the continuous improvement in supply chain competitiveness, the management take
actions in cases of any change in supply chain environment and risk levels. In particular, costs and premium
freights related to the critical sub-network are continuously monitored. The supply chain manager can revise
the mitigation plan according to the changes in critical sub-network performance. In a stationary supply chain
environment, the manager will mainly focus on supply chain efficiency. In this case, the effects of low safety
stock levels or low supplier flexibility on supply chain performance should be analysed. In turbulent
environments, the manager will focus on keeping premium freights under control. Moreover, risk identification
and mitigation phases can be reiterated in case of a major change in supply chain environment. Furthermore,
the criteria for critical sub-network identification can be reconsidered.
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
631
4. Application of the Proposed Framework
The proposed framework is applied to a supply chain of a European automotive company. The supply chain
consists of 752 suppliers, 12 plants and 55 customers. The plants assemble the materials into a semi-finished
product specialized to the customer. The customers are just-in-time automobile manufacturers. Therefore,
backlogging is not allowed in the supply chain. In case of a shortage risk, a premium freight is requested.
The supply chain under concern operates six days in a week and 48 weeks in a year. Customers share their
demand forecast with the plants in weekly basis. The plants adjust the customer demand forecast information
by a demand forecast adjustment factor. In particular, customer demand forecast information is adjusted by
multiplying it with demand forecast adjustment factor. The plants use the adjusted demand forecast information
in developing their production plan and determining the order sizes to be placed to their suppliers.
The plants use a periodic order-up-to policy for the materials’ inventory. At the beginning of each week, the
weekly requirements for each material are calculated. Then, if it is needed, an order is placed. The order
quantities are determined by considering safety stock level, transportation lead time and average daily
consumption of material, as well as, quantity flexibility limits of the suppliers. The orders are processed by
suppliers immediately and are delivered after a transportation lead time. Shipments and deliveries are made
only in working days. Production and storage processes of the suppliers are not considered in this study.
Inventory position of each plant is reviewed on daily basis. If inventory position of a material is below of the
safety stock level and it is not the regular ordering day, an inbound premium freight must be requested from
the supplier of the material. At the end of each week, customer demand is realized and filled from the inventory.
The customer demand cannot be backordered. If it is not possible to deliver on time, final products are delivered
to the customer by an outbound premium freight. The premium freights are delivered at the next day of the
shipment.
The required data for the analysis are obtained from the plan for every part spreadsheets which are used by the
plants for production planning. The demand forecast adjustment factor for each plant is 1.07. The plants holds
3.5 days of safety stock for each material. Supplier flexibility is presented as a percentage in the quantity
flexibility contracts. In these contracts, % flexibility means that the order quantity of a plant can be %
below or above of the contracted quantity. Currently, quantity flexibility of the suppliers is 50%. The quantity
flexibility can be increased by making a new contract. However, the contracted unit price will be higher in
case of a higher quantity flexibility. By consulting a supply chain manager, we model the relationship between
supplier flexibility and unit price as in the following.
(4.1)
where is the current unit price under current quantity flexibility (), and
is the new unit price which is specified for quantity flexibility.
In subsequent sections, identification and modeling of the critical sub-networks through material-level and
product level risk management considerations are presented.
4.1 Material-Level Risk Management
This section focuses on the supply chain risks related to materials. Hence, the materials are investigated in
terms of their criticality by the proposed risk identification procedure. As the inbound supply chain inventory
consists of 7300 materials consumed by 12 plants, it is unreasonable to assess the risks related to all materials.
Therefore, the materials that have a considerable share on annual inbound premium freight costs are identified
by using Pareto principle. As a result, 37 out of 7300 materials presenting 80% of annual inbound premium
cost are selected for risk identification.
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
632
4.1.1 Risk Identification
The inbound risks of the focal supply chain are identified as supplier delivery delay and delivery quantity
loss. The inbound risks are modelled by their occurrence and severity values. The fractions of late and under-
shipped deliveries are obtained by historical data and used as the occurrence values of delivery delay and
delivery quantity loss, respectively. To quantify the severity values of inbound risks, delivery lateness and
quantity loss data are obtained from historical order records. The historical data are fitted to a number of
probability distributions such as normal, lognormal, exponential, Weibull and gamma distributions to obtain
the risk models.
The outbound risk is identified as variability in the material consumption. However, since a material type can
be used in production of hundreds of different product types, it is impractical to model material consumption
variability based on customer demand variability. Therefore, material consumption variability is assumed to
follow a uniform distribution varying between 50% and 150% of average daily consumption in parallel with
the supply chain manager’s opinion.
Afterwards, the focal supply chain is decomposed into a critical sub-network by considering inbound supply
chain risk performance. The criteria for critical sub-network identification are determined by the supply chain
manager as number of inbound premium freights in previous year, monetary value of inbound premium
freights in previous year and average weekly consumption of materials. However, these criteria are assessed
by the plants with different ratings. Therefore, multiple decision matrices are formed. As stated previously,
these decision matrices should be merged into unique decision matrix by using a group decision making
approach. In this study, the decision matrices are merged by averaging the ratings of the plants. In the TOPSIS,
the decision matrices are normalized into [0,1]. Hence, the positive ideal solution () is one while the negative
ideal solution () is zero for all criteria. In the calculation of overall criticality index, the criteria weights are
obtained as 0.33, 0.33, and 0.34 for the number of premium freights, the monetary value of premium freights
and the average weekly consumption, respectively. The overall criticality indices calculated by using TOPSIS
are presented in Table 1. As it can be seen from the table, the most critical material is M1. The critical sub-
network related to material M1 is presented in Figure 3.
Table 1
Overall criticality indices for materials
Material
Material
Material
Material
M1
0.14
M9
0.03
M8
0.02
M21
0.01
M5
0.11
M24
0.03
M36
0.02
M31
0.01
M16
0.08
M34
0.03
M18
0.02
M12
0.01
M2
0.05
M6
0.03
M32
0.02
M28
0.01
M3
0.05
M29
0.03
M33
0.02
M20
0.01
M14
0.05
M17
0.02
M30
0.02
M26
0.00
M11
0.04
M19
0.02
M10
0.01
M35
0.00
M7
0.04
M22
0.02
M27
0.01
M13
0.04
M4
0.02
M23
0.01
M25
0.03
M15
0.02
M37
0.01
4.1.2 Risk Mitigation
In this phase, a simulation model of the critical sub-network is developed. The inbound risks affecting the
critical sub-network are modelled by the aforementioned risk quantification approach. The delivery loss
probability is calculated as 0.05. To determine the probability distribution of the delivery loss quantity, the
historical delivery loss data are fitted to a number of distributions in MINITAB statistical software. The best
fitted distribution is a lognormal distribution with parameters 0.52 and 0.23. The delivery delay probability is
0.06. The delivery delay time distribution is specified as a lognormal (1.43, 0.46). By consulting the supply
chain manager, material consumption variability distribution is specified as a uniform distribution varying
between 50% and 150% of average daily consumption.
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
633
The performance measures are annual holding cost of supply chain and inbound premium freight ratio. Annual
holding cost of supply chain involves annual holding costs of the plants associated with the critical material.
Inbound premium ratio is the ratio of total amount of premium freights to total amount of orders associated
with the critical material in the supply chain.
Figure 3. The critical sub-network identified in case of material-level risk management
The critical sub-network model is simulated for 104 weeks. The warm-up period is determined as two weeks
by inspecting the variability in orders. Furthermore, required number of replications are determined by
analysing the simulation outputs. In particular, a number of confidence intervals are calculated for annual
holding cost and inbound premium freight ratio at the end of each replication. Until the confidence intervals
become narrow enough, the replications proceed. The required half-length for the confidence intervals are
0.15. As a result, 15 replications are found to be sufficient to predict the performance measures within the
predefined error rate.
To analyse the effects of risk mitigation strategies, a full factorial experimental design (Montgomery, 2008) is
developed by considering five factor levels for safety stock and supplier flexibility (see Table 2). The response
variables of the design are the annual holding cost and the premium freight ratio. The values for the response
variables are obtained by the simulation model. According to the results of ANOVA, the effects of safety stock
and supplier flexibility are significant on both annual holding cost and premium freight performances. To
determine the best factor levels yielding minimum annual holding cost and premium freight ratio, MINITAB
Response Optimizer tool is used. The best factor levels are determined as 4.5 days for safety stock and 30%
for supplier flexibility. Comparison of the performances corresponding to the best factor levels and current
factor levels is presented in Table 3. As one can see from the table, the new factor levels reduce the annual
holding cost by 8% and the premium freight ratio by 3%.
Table 2
Factor levels considered in case of material-level risk management
Factors
Levels
Safety stock
2.5
3
3.5*
4
4.5
Flexibility
30%
40%
50%*
60%
70%
*The current levels used in the supply chain
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
634
Table 3
Performance comparison in case of material-level risk management
Safety
Stock
Supplier
Flexibility
Annual
Holding Cost
Inbound Premium
Freight Ratio
New factor levels
4.5
30%
€4015
0.19
Current factor levels
3.5
50%
€4349
0.20
4.2 Product-Level Risk Management
In this section, the products are investigated in terms of their criticality by the proposed risk assessment
procedure. Since there exist 2700 different types of products produced by the plants, it is unreasonable to deal
with the risks associated with all products. Hence, the products that have a considerable share on annual
outbound premium freight costs are identified by using Pareto principle. As a result, 13 out of 2700 products
presenting 80% of annual outbound premium cost are selected for risk identification.
4.2.1 Risk Identification
In the product-level risk management case, the inbound risks are identified as supplier delivery delay and
delivery loss. The occurrence and severity values are quantified by the probability distributions derived from
past order data as the in material-level risk management case. Outbound risks are identified as the customer
demand variability and the deviation of actual customer demand from the shared demand information. These
risks are modelled as ordinary variabilities that occur every week and only severity values of them are
modelled. To model outbound risks, the demand data are obtained from the historical customer demand
records.
Then, the focal supply chain is decomposed into a critical sub-network by considering outbound supply chain
risk performance. By consulting the supply chain manager, the criteria for critical sub-network identification
are determined as the number of outbound premium freights in previous year, monetary value of the outbound
premium freights in previous year and average annual sales of the products. As in the material-level risk
management case, the decision matrices are normalized into [0,1]. Therefore, the positive ideal solution is one
and the negative ideal solution is zero for all criteria. The entropy weights are obtained as 0.34, 0.34, and 0.32
for the number of outbound premium freights, the monetary value of premium freights and the average annual
sales, respectively. The overall criticality indices calculated by using TOPSIS method are presented in Table
4. As can be inferred from the table, the most critical product is PR1. The critical sub-network related to PR1
is presented in Figure 4.
4.2.2 Risk Mitigation
In this phase, a simulation model of the critical sub-network is developed. The bill of material information
related to PR1 is reported in Table 5.
The inbound risks affecting the critical sub-network are modelled by the aforementioned risk quantification
approach. The probability distributions that are best fitted to historical data are given in Table 6. The best fitted
distributions are normal and Weibull distributions for delivery loss quantity and delivery delay, respectively.
The customer demand is modelled by the demand forecast and the deviation of actual customer demand from
the demand forecast. The best fitted probability distributions for the demand forecast and the deviation are
Lognormal(6.27, 0.51) and Lognormal(0.07, 0.21), respectively.
The performance measures are annual holding cost of supply chain, inbound premium freight ratio, and
outbound premium freight ratio. Annual holding cost and inbound premium freight ratio are calculated as in
Section 4.1.2. Outbound premium freight ratio is the ratio of total amount of the premium freights related to
the products to the total amount of products sold.
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
635
Table 4
Overall criticality indices for the products
Products
PR1
0.18
PR11
0.14
PR6
0.14
PR2
0.12
PR3
0.10
PR5
0.10
PR4
0.08
PR12
0.07
PR7
0.05
PR13
0.05
PR8
0.04
PR10
0.03
PR9
0.03
Figure 4. The critical sub-network identified in case of product-level risk management
Table 5
Bill of material for PR1
Material
Supplier
Quantity
M2
S2
1
M22
S2
3
M71
S7
3
M73
S7
3
M76
S7
1
M79
S7
3
Table 6
The inbound risk model in case of product-level risk management
Delivery Loss
Delivery Delay
Suppliers
Probability
Quantity distribution
Probability
Time distribution
S2
0.05
Normal(0.45,0.17)
0.09
Weibull(1.65,1.61)
S7
0.03
Normal(0.51,0.25)
0.07
Weibull(1.20,2.30)
The warm-up period length and the number of replications are determined in the same manner described in
Section 4.1.2. The critical sub-network model is simulated for 104 weeks and 15 replications. The warm-up
period is specified as one week. To analyse the effects of risk mitigation strategies, a full factorial experimental
design is developed for five factor levels for demand forecast adjustment factor, safety stock and supplier
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
636
flexibility. (see Table 7). The response variables are the annual holding cost, inbound and outbound premium
freight ratios. The response values are obtained from the simulation model.
Table 7
Factor levels considered in the experiment
Factors
Levels
Demand adj.
1
1.03
1.07*
1.11
1.15
Safety stock
2.5
3
3.5*
4
4.5
Flexibility
30%
40%
50%*
60%
70%
*The current levels used in the supply chain
According to the ANOVA results, all the factors have significant effect on the holding cost and the outbound
premium freight ratio. However, only the supplier flexibility affects the inbound premium freight ratio
significantly. The best factor levels yielding minimum annual holding cost, inbound and outbound premium
freight ratios are determined via MINITAB Response Optimizer Tool. The best factor levels are 1.15, 2.5 and
30% for demand forecast adjustment factor, safety stock and supplier flexibility, respectively. Comparison of
the performances corresponding to these factor levels and current factor levels is presented in Table 8. The
results reveal that the new factor levels reduce annual holding cost by 10%, inbound premium freight ratio by
26%, and outbound premium freight ratio by 28%.
Table 8
Performance comparison of in case of product-level risk management
Demand
Adj. Fac.
Safety
Stock
Supplier
Flexibility
Annual
Holding Cost
Inbound Pre-
mium Freight
Ratio
Outbound Pre-
mium Freight
Ratio
New factor levels
1.15
2.5
30%
€20254
0.07
0.04
Current factor levels
1.07
3.5
50%
€22425
0.10
0.06
5. Discussion and Managerial Implications
According to the results, the proposed framework is capable of ensuring a substantial improvement in terms
of holding cost and premium freight performances. In this application, emphasizing on redundancy strategies
improves the supply chain risk performance in both holding cost and premium freight aspects. In Figure 5, the
main effects of safety stock and supplier flexibility levels on holding cost and premium freight ratio are given
for the material-level risk management case. The ANOVA results shows that the effects of supplier flexibility
on annual holding cost and premium freight ratio are significant. Moreover, safety stock level has a significant
effect on premium freight ratio. In this case, higher supplier flexibility levels yield lower premium freight ratio,
but increase holding cost. On the other hand, main effects of safety stock levels on premium freight ratio is
nonlinear where the current safety stock level (3.5 days) incurs the worst premium freight performance. Thus,
a compromise solution can be obtained among two objectives by decreasing supplier flexibility level and
increasing safety stock level. In accordance with this result, the proposed framework improves both holding
cost and premium freight ratio by reducing supplier flexibility and increasing safety stock levels (see Table 3).
In Figure 6, main effects of demand forecast adjustment factor, safety stock, and supplier flexibility levels on
supply chain performance are illustrated for product-level risk management case. According to the ANOVA
results, demand forecast adjustment factor, safety stock and supplier flexibility have significant effect on
annual holding cost and outbound premium freight ratio. Moreover, supplier flexibility has a significant effect
on inbound premium freight ratio. As one can infer from the figure, demand forecast adjustment factor has
lower effect on holding cost and higher effect on outbound premium freight ratio than the other parameters.
Therefore, to reduce outbound premium freight ratio, demand forecast adjustment factor can be increased.
Moreover, to reduce holding cost supplier flexibility can be reduced by considering its significant relationship
with demand forecast adjustment factor in terms of outbound premium freight ratio. Furthermore, we can infer
from the figure that safety stock has relatively low effect on outbound premium freight ratio. Accordingly, the
proposed framework ensures better supply chain performance in three objectives by increasing demand
forecast adjustment factor and decreasing safety stock and supplier flexibility levels (Table 8).
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
637
Figure 5. Main effects of parameter levels on supply chain performance in material-level risk management
case
In both cases, the framework suggests lower supplier flexibility levels than the current levels. In material-level
risk management case, safety stock levels are increased. In the product-level risk management case, demand
forecast adjustment factor is increased. Consequently, redundancy strategies are preferred rather than
flexibility strategy in this application. Therefore, the managers of the focal supply chain should put more
emphasis on redundancy strategies (high demand forecast adjustment factor and safety stock levels).
Figure 6. Main effects of parameter levels on supply chain performance in product-level risk management case
Supply chain managers can use the proposed framework in cases of any changes in supply chain environment,
competition strategy, and new supplier contracts. Although the objectives have equal importance in this
application, the proposed framework provides the flexibility in evaluating risk mitigation plans by considering
different weights for the objectives. In a stable supply chain environment, the manager can give more
importance to holding costs. In turbulent supply chain environment, the manager will mainly focus on supply
chain risks and reduce premium freights. Moreover, in case of a change in competition strategy, the manager
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
638
may consider reducing holding costs to gain cost advantage, or focus on supply chain risks to ensure customer
satisfaction. Furthermore, the proposed framework will be beneficial in making new supplier contracts, since
it considers both cost and resilience objectives.
The proposed framework provides a comprehensive decision support since it involves both material and
product-level risk analyses through the preference of the decision maker. Additionally, it considers redundancy
and flexibility strategies in combined manner to improve supply chain risk performance efficiently and
effectively. Moreover, it measures the supply chain vulnerability by premium freight ratio. Furthermore, due
to its flexible and convenient structure, it can be applied to various supply chain structures.
6. Conclusion
This study proposes an integrated risk management framework for global supply chains. In the risk
identification phase of the proposed framework, global supply chain is decomposed into material-level or
product-level critical sub-networks according to preference of the manager. Consequently, the proposed
framework is applicable for both material and product-level risk analyses. This provides the flexibility in
choosing the focus of SCRM through the manager’s performance objectives. Additionally, the proposed
framework enables managers in combining redundancy and flexibility strategies to ensure both effectiveness
and efficiency objectives in SCRM. In this study, an application of the proposed framework to an automobile
supply chain is presented. The results of both material and product level analyses show that the proposed
framework improves the supply chain performance.
A limitation of this study is the assumption of the identical safety stock and supplier flexibility levels
throughout the supply chain. These parameters may take different values for each material and supplier.
However, this increases the complexity of the problem. Consequently, finding the best parameter levels by
using an experimental design approach become challenging. Furthermore, analysis of supply chain risk drivers,
and considering rare and severe adverse events in the proposed framework are possible future research
directions.
Author Contributions
Mualla Gonca Avci: Conceptualization, Methodology, Software, Validation, Data curation, Writing
Conflicts of Interest
The authors declare no conflict of interest.
References
Aqlan, F., & Lam, S. S. (2015a). A fuzzy-based integrated framework for supply chain risk assessment.
International Journal of Production Economics, 161, 54-63. doi:
https://doi.org/10.1016/j.ijpe.2014.11.013
Aqlan, F., & Lam, S. S. (2015b). Supply chain risk modelling and mitigation. International Journal of
Production Research, 53(18), 5640-5656. doi: https://doi.org/10.1080/00207543.2015.1047975
Bueno-Solano, A., & Cedillo-Campos, M. G. (2014). Dynamic impact on global supply chains performance
of disruptions propagation produced by terrorist acts. Transportation Research Part E: Logistics and
Transportation Review, 61, 1-12. doi: https://doi.org/10.1016/j.tre.2013.09.005
Cardoso, S. R., Paula Barbosa-Póvoa, A., Relvas, S., & Novais, A. Q. (2015). Resilience metrics in the
assessment of complex supply-chains performance operating under demand uncertainty. Omega, 56, 53-
73. doi: https://doi.org/10.1016/j.omega.2015.03.008
Carvalho, H., Barroso, A. P., Machado, V. H., Azevedo, S., & Cruz-Machado, V. (2012). Supply chain
redesign for resilience using simulation. Computers & Industrial Engineering, 62(1), 329-341. doi:
https://doi.org/10.1016/j.cie.2011.10.003
Chaudhuri, A., Mohanty, B. K., & Singh, K. N. (2013). Supply chain risk assessment during new product
development: a group decision making approach using numeric and linguistic data. International
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
639
Journal of Production Research, 51(10), 2790-2804. doi:
https://doi.org/10.1080/00207543.2012.654922
Chen, P.-S., & Wu, M.-T. (2013). A modified failure mode and effects analysis method for supplier selection
problems in the supply chain risk environment: A case study. Computers & Industrial Engineering,
66(4), 634-642. doi: https://doi.org/10.1016/j.cie.2013.09.018
Chopra, S., & Sodhi, M. S. (2004). Managing Risk To Avoid Supply-Chain Breakdown. MIT Sloan
management review, 46(1), 53-61.
Christopher, M., & Peck, H. (2004). Building the Resilient Supply Chain. International Journal of Logistics
Management, 15(2), 1 - 14. doi: https://doi.org/10.1108/09574090410700275
Deng, H., Yeh, C. H., & Willis, R. J. (2000). Inter-company comparison using modified TOPSIS with objective
weights. Computers & Operations Research, 27(10), 963-973. doi: https://doi.org/10.1016/S0305-
0548(99)00069-6
Colicchia, C., & Strozzi, F. (2012). Supply chain risk management: a new methodology for a systematic
literature review. Supply Chain Management: An International Journal, 17(4), 403-418. doi:
https://doi.org/10.1108/13598541211246558
Ghadge, A., Dani, S., Chester, M., & Kalawsky, R. (2013). A systems approach for modelling supply chain
risks. Supply Chain Management: An International Journal, 18(5), 523-538. doi:
https://doi.org/10.1108/SCM-11-2012-0366
Giannakis, M., & Louis, M. (2011). A multi-agent based framework for supply chain risk management.
Journal of Purchasing and Supply Management, 17(1), 23-31. doi:
https://doi.org/10.1016/j.pursup.2010.05.001
Govindan, K., & Jepsen, M. B. (2015). Supplier risk assessment based on trapezoidal intuitionistic fuzzy
numbers and ELECTRE TRI-C: a case illustration involving service suppliers. Journal of the
Operational Research Society, 67(2), 339-376. doi: https://doi.org/10.1057/jors.2015.51
Guertler, B., & Spinler, S. (2015). When does operational risk cause supply chain enterprises to tip? A
simulation of intra-organizational dynamics. Omega, 57, 54-69. doi:
https://doi.org/10.1016/j.omega.2015.03.005
Heckmann, I., Comes, T., & Nickel, S. (2015). A critical review on supply chain risk – Definition, measure
and modeling. Omega, 52, 119-132. doi: https://doi.org/10.1016/j.omega.2014.10.004
Ho, W., Zheng, T., Yildiz, H., & Talluri, S. (2015). Supply chain risk management: a literature review.
International Journal of Production Research, 53(16), 5031-5069. doi:
https://doi.org/10.1080/00207543.2015.1030467
Hwang, C. L., & Yoon, K. P. (1981). Multiple attribute decision making: Methods and applications: Springer-
Verlag, New York.
Kara, M. E., Fırat, S. Ü. O. & Ghadge, A. (2020). A data mining-based framework for supply chain risk
management. Computers & Industrial Engineering, 139, 105570. doi:
https://doi.org/10.1016/j.cie.2018.12.017
Klibi, W., & Martel, A. (2012). Scenario-based Supply Chain Network risk modeling. European Journal of
Operational Research, 223(3), 644-658. doi: https://doi.org/10.1016/j.ejor.2012.06.027
Montgomery, D. C. (2008). Design and analysis of experiments: John Wiley & Sons.
Oliveira, U. R., de Almeida Muniz, M., Anaia, L. A., & Rocha, H. M. (2022). Medication supply chain risk
management for a brazilian home care provider: A business sustainability study. Cleaner Logistics and
Supply Chain, 3, 100018. doi: https://doi.org/10.1016/j.clscn.2021.100018
Oliveira, J. B., Jin, M., Lima, R. S., Kobza, J. E., & Montevechi, J. A. B. (2019). The role of simulation and
optimization methods in supply chain risk management: Performance and review standpoints.
Simulation Modelling Practice and Theory, 92, 17-44. doi: https://doi.org/10.1016/j.simpat.2018.11.007
Pournader, M., Kach, A., & Talluri, S. (2020). A review of the existing and emerging topics in the supply
chain risk management literature. Decision Sciences, 51(4), 867-919. doi:
https://doi.org/10.1111/deci.12470
Rajesh, R., & Ravi, V. (2015). Modeling enablers of supply chain risk mitigation in electronic supply chains:
A Grey–DEMATEL approach. Computers & Industrial Engineering, 87, 126-139. doi:
https://doi.org/10.1016/j.cie.2015.04.028
Journal of Advanced Research in Natural and Applied Sciences 2022, Vol. 8, Issue 4, Pages: 624-640
640
Rangel, D. A., de Oliveira, T. K., & Leite, M. S. A. (2014). Supply chain risk classification: discussion and
proposal. International Journal of Production Research, 1-20. doi:
https://doi.org/10.1080/00207543.2014.910620
Samvedi, A., Jain, V., & Chan, F. T. S. (2013). Quantifying risks in a supply chain through integration of fuzzy
AHP and fuzzy TOPSIS. International Journal of Production Research, 51(8), 2433-2442. doi:
https://doi.org/10.1080/00207543.2012.741330
Schmitt, A. J., & Singh, M. (2012). A quantitative analysis of disruption risk in a multi-echelon supply chain.
International Journal of Production Economics, 139(1), 22-32. doi:
https://doi.org/10.1016/j.ijpe.2012.01.004
Sheffi, Y. (2005). A Supply Chain View of the Resilient Enterprise. MIT Sloan management review, Fall 2005,
41-48.
Simchi-Levi, D., Schmidt, W., Wei, Y., Zhang, P. Y., Combs, K., Ge, Y., . . . Zhang, D. (2015). Identifying
Risks and Mitigating Disruptions in the Automotive Supply Chain. Interfaces, 45(5), 375-390. doi:
https://doi.org/10.1287/inte.2015.0804
Singhal, P., Agarwal, G., & Mittal, M. L. (2011). Supply chain risk management: review, classification and
future research directions. International Journal of Business Science and Applied Management, 6(3),
15-42.
Sodhi, M. S., Son, B.-G., & Tang, C. S. (2012). Researchers' Perspectives on Supply Chain Risk Management.
Production and Operations Management, 21(1), 1-13. doi: https://doi.org/10.1111/j.1937-
5956.2011.01251.x
Talukder, B., Agnusdei, G. P., Hipel, K. W., & Dubé, L. (2021). Multi-indicator supply chain management
framework for food convergent innovation in the dairy business. Sustainable Futures, 3, 100045. doi:
https://doi.org/10.1016/j.sftr.2021.100045
Tang, O., & Nurmaya Musa, S. (2011). Identifying risk issues and research advancements in supply chain risk
management. International Journal of Production Economics, 133(1), 25-34. doi:
https://doi.org/10.1016/j.ijpe.2010.06.013
Tummala, R., & Schoenherr, T. (2011). Assessing and managing risks using the Supply Chain Risk
Management Process (SCRMP). Supply Chain Management: An International Journal, 16(6), 474 -
483. doi: https://doi.org/10.1108/13598541111171165
Wagner, S. M., Mizgier, K. J., & Arnez, P. (2014). Disruptions in tightly coupled supply chain networks: the
case of the US offshore oil industry. Production Planning & Control, 25(6), 494-508. doi:
https://doi.org/10.1080/09537287.2012.705355
Wang, X., Chan, H. K., Yee, R. W. Y., & Diaz-Rainey, I. (2012). A two-stage fuzzy-AHP model for risk
assessment of implementing green initiatives in the fashion supply chain. International Journal of
Production Economics, 135(2), 595-606. doi: https://doi.org/10.1016/j.ijpe.2011.03.021
