Supply Chain Risk Assessment Applying System Dynamics Approach
Case Study: Apparel Industry
Marzieh Mehrjoo and Zbigniew J. Pasek
Department of Industrial and Manufacturing Systems Engineering, University of Windsor, Windsor, Canada
Keywords: Supply Chain, Risk Assessment, System Dynamics, Simulation, Apparel Industry.
Abstract: A remarkable increase in the demand and supply uncertainty, as the primary sources of supply chain risk
and other sources such as: capacity constraints, supply variability, parts quality problems, long lead times,
war and natural disasters have increased the necessity of assessing and managing the risk in the supply
chain. The purpose of this study is to investigate the impact of two categories of risk, demand uncertainty
and delays, on the performance of an apparel supply chain. A system dynamics approach was used to study
the behavior and relationships within the supply chain of this industry. The proposed model facilitates the
study and identification of the critical components of the supply chain. In addition, the model provides a
tool to generate multiple business scenarios for effective decision making.
1 INTRODUCTION
Uncertainty in the demand for products is the
primary source of risk in the supply chain. Several
interdependent factors such as higher product
variety, shorter product life cycles, increased
customer expectations, more complex and longer
supply chains, and more global competitions have
increased this uncertainty considerably in the recent
years. Moreover, capacity constraints, supply
variability, parts quality problems, long lead times,
and manufacturing yields besides disruptions due to
war and natural disasters are some other sources of
risks in the supply chain (Sheffi and Rice, 2005).
Therefore, it is essential for companies to understand
supply chain interdependencies, identify potential
risk factors, their likelihood and consequences
(Tummala and Schoenherr, 2011). They need to
develop plans for disruptions and contingency plans
to decrease the likelihood of supply chain risks.
In the apparel industry, market demand is highly
volatile and product life cycles are short. Low
predictability and high level of impulse purchase are
other characteristics of this market (Carugati et al.,
2008). All the previously mentioned factors increase
the importance of risk assessment for the supply
chain of these products. Barlas and Aksogan (1997)
built a system dynamics simulation model for the
textile and apparel pipeline which consisted of
wholesaler and retailer levels. They studied the
effects of product diversity and quick response order
strategies on customer demand, possible stockouts
and inventory levels.The purpose of this study is to
investigate the impact of two categories of risks -
demand uncertainty and delays - on the performance
of an apparel supply chain through a system
dynamics approach.
2 MODEL
Industrial dynamics (Forrester, 1961) introduced a
methodology for the simulation of dynamic models,
which is the origin of system dynamics (Sterman,
2000). Extensive research in various fields, natural
and social sciences, has been conducted using
system dynamics. System dynamics is one of the
best methods for analyzing complex systems
(Campuzano and Mula, 2011). In this study, the
system dynamics software, Vensim®, was used as a
tool to build the supply chain model. The model (see
Fig. 5) has three primary members: a manufacturer,
a distributor and a retailer.
At the retailing level for an apparel store, when
the customer’s need is not satisfied, the customer
usually leaves and does not wait for his/her need to
be fullfilled. Therefore, the constructed model has
considered the orders not delivered on time as the
lost sale at this stage. If the warehouse contains the
sufficient amount of products to meet the demand at
343
Mehrjoo M. and Pasek Z..
Supply Chain Risk Assessment Applying System Dynamics Approach - Case Study: Apparel Industry.
DOI: 10.5220/0004276601450148
In Proceedings of the 2nd International Conference on Operations Research and Enterprise Systems (ICORES-2013), pages 145-148
ISBN: 978-989-8565-40-2
Copyright
c
2013 SCITEPRESS (Science and Technology Publications, Lda.)
the time, the order is delivered. Otherwise, the final
customer’s demand is transferred to increase the
“Retailer stockout” variable. Based on the inventory
position, forecasted demand and lead time of this
stage, the retailer’s replenishment orders will be sent
to the distributor through the “Orders to distributor”
variable. The output of “On order products
(retailer)” is “Products delivered to retailer” which is
affected by the lead time and introduces a delay into
the arrival of products. The delay is considered to be
pure and not exponential, that means that the arrival
of products at the warehouse happens exactly after
the period defined in the “Retailer lead time”
variable.
At the distributor level, if the orders are not met
by the required date, they will be served when the
distributor has enough stock available. The proposed
model considered the orders not delivered on time as
backlogged orders and they were included in the
daily firm orders. The flow of information and
material at the distributor level undergoes the same
transformations as the previous level except the case
of backlogged orders that do not exist in the retailing
level.
The service and delivery policies for backlogged
or delayed orders at the manufacturer level follow
the same formulation as the distributor level. The
manufacturer has a predefined daily capacity, so it
can only manufacture the amount of units the factory
is capable of.
The demand forecasts are calculated during each
period based on a simple exponential smoothing
technique.
In the proposed model, the total cost is calculated
based on the unit cost, the amount of products
delivered including delivered backlogged orders,
products manufactured for the manufacturer level
and products purchased for the other two stages. It is
assumed that the distributor is the member
responsible for all the transportations taking place in
the supply chain. Therefore, the variables of
“transportation revenue” and the “transportation
cost” are included in this level. It should be
mentioned that a variable to calculate the total
revenue of each stage is not defined separately, but it
is calculated inside the formula of the profit variable.
In order to test the validity of the model, three
different tests including direct extreme conditions
test, dimentional consistency test and direct structure
test, by comparing the model equations with
available knowledge in the literature, were
conducted.
The main characteristics of the model including
model parameters and assumptions are as follows:
It is possible to serve only one part of the
order when the whole order is not available.
Inventory management is performed applying
inventory review policy.
The raw materials used for manufacturing are
considered to be available all the time.
Simulation takes place over 365 periods.
The stock of the initial inventory for the
manufacturer level is 10 units and for both,
distributor and retailer, are 5 units.
The manufacturing capacity is 25 units per
period.
The manufacturing lead time is 8 periods and
the manufacturing lead time to the distributor
and then to the retailer are 2 and 1 periods,
respectively.
The adjust factor for forecasting is equal to 2.
The pattern selected for the number of
customers entering the retail store per period
corresponds to a normal distribution with the
μ = 25 and σ = 9.
Since a customer who enters the store may
leave without buying an item, a binomial
distribution was selected to calculate the
probability that a customer will prefer a SKU
in the retail store.
Actual customer’s demand is calculated by
multiplication of the number of customers
entering the retail store and the probability
that a customer will buy an item after
entering the store.
Number of color, style and size varieties of
products are 5, 12 and 10, respectively.
3 RISK ASSESSMENT
As mentioned previously, this paper studies the
effect of demand variability and the risk of delay on
the supply chain performance of an apparel industry.
The cumulative cost in each stage of the supply
chain is used as the performance measure.
3.1 Risk of Delay
In order to investigate the impact of delay, four
different scenarios have been considered. In the
Main scenario there is no increase in the lead time of
any of the stages; in LT1 there are 6 periods of time
increase only in the lead time of manufacturer; in
LT2 there are 5 periods of time increase only in the
lead time of distributor; and in LT3 there are 2 units
of time increase only in the lead time of the retailer.
Figures 1 to 3 depict the results of the simulation. It
ICORES2013-InternationalConferenceonOperationsResearchandEnterpriseSystems
344
can be seen that the Main scenario, with no delay in
the model, has the least cost in all the three stages of
the supply chain. Also, delay in the lead time of
manufacturer level, LT1, causes the highest cost
increase in the manufacturing and distribution
stages. The costs in the retailing stage are mainly
sensitive to the delays in the same stage, retailing
(not shown).
Figure 1: Effect of delay in lead time on manufacturer
cumulative total cost.
Figure 2: Effect of delay in lead time on distributor
cumulative total cost.
3.2 Risk of Demand Variability
Three scenarios were defined to compare the
performance of the supply chain under various
variabilities of demand patterns. The number of
customers entering the retailing store (potential
customer demand) follows a normal distribution
with the following characteristics
Main: Range = 40 and σ = 9,
Dem 1: Range = 40 and σ = 25.
Dem 2: Range = 90 and σ = 9.
From Fig. 3, it can be interpreted that increasing the
variation of demand has a higher impact on the
manufactuer and distributor compared to the retailer,
considerably increasing their costs. In addition, these
two stages perform worse when the data related to
the demand comes from an interval with longer
width (Fig. 4). In the designed supply chain, the
retailer stage shows the least sensitivity to the
variation and uncertainty in demand.
Figure 3: Effect of demand variability on manufacturer
cumulative total cost.
Figure 4: Effect of demand variability on distributor
cumulative total cost.
4 SUMMARY
This paper presented a system dynamics model that
can be used to study and observe the processes and
relationships in the supply chain of an apparel
industry. It is also useful as a high-level tool to
analyze the impact of different types of risks
associated with the supply chain of this industry.
This study investigated the impact of demand
uncertainty and risk of delay on the supply chain
performance. The main limitation of the model is the
data used as the inputs. Due to the inability to obtain
more accurate industry-specific data, the absolute
numbers that this model presented are used for
comparative analysis of different scenarios.
Nevertheless, the model helps to study the
relationships between supply chain stages and to
analyze the effect of changing values for the
variables of the model.
0
10000
20000
30000
40000
50000
0 50 100 150 200 250 300 350
CumulativeTotalCost(Mfg.)[$]
Time[periods]
LT1
LT3
LT2
main
0
50000
100000
150000
200000
250000
0 50 100 150 200 250 300 350
CumulativeTotalCost(Distr.)[$]
Time [periods]
LT1
LT3
LT2
main
0
20000
40000
60000
80000
100000
120000
140000
160000
180000
0 50 100 150 200 250 300 350
CumulativeTotalCost(Distr.) [$]
Dem1
Dem2
main
Time [periods]
0
5000
10000
15000
20000
25000
30000
35000
40000
0 50 100 150 200 250 300 350
CumulativeTotalCost(Mfg.)[$]
Dem1
Dem2
main
Time [days]
SupplyChainRiskAssessmentApplyingSystemDynamicsApproach-CaseStudy:ApparelIndustry
345
REFERENCES
Barlas, Y., Aksogan, A., 1997. Product Diversification and
Quick Response Order Strategies in Supply Chain
Management. Available from
http://ieiris.cc.boun.edu.tr/faculty/barlas.
Campuzano, F. and Mula, J., 2011. Supply Chain
Simulation: A System Dynamics Approach for
Improving Performance. Springer London Dordrecht
Heidelberg New York.
Carugati, A., Liao, R., Smith, P., 2008. Speed-to-Fashion:
Managing Global Supply Chain in Zara. Proceedings
of the IEEE ICMIT, pp. 1494-1499.
Choi, T.Y. Krause, D.R., 2006. The supply base and its
complexity: implications for transaction costs, risks,
responsiveness, and innovation. Journal of Operations
Management, Vol. 24 No. 5, pp. 637-52.
Forrester, J.W., 1961. Industrial Dynamics. MIT Press and
Wiley, New York.
Sheffi, Y., Rice Jr., J.B., 2005. A Supply chain view of the
resilient enterprise. MIT Sloan Management Review,
Vol. 47 No. 1., pp. 41-48.
Sterman, J.D., 2000. Business Dynamics: Systems
Thinking and Modeling for a Complex World.
McGraw-Hill Higher Education, New York.
Tummala, R., Schoenherr, T., 2011. Assessing and
managing risks using the Supply Chain Risk
Management Process (SCRMP). Supply Chain
Management: An International Journal, Vol. 16 No. 6,
pp. 474–483.
Zsidisin, G.A., Ellram, L.M., Carter, J.R. Cavinato, J.L.,
2004. An analysis of supply risk assessment
techniques. International Journal of Physical
Distribution & Logistics Management, Vol. 34 No. 5,
pp. 397-409.
Zsidisin, G.A., Panelli, A., Upton, R. 2000. Purchasing
organization involvement in risk assessments,
contingency plans, and risk management: an
exploratory study. Supply Chain Management: An
International Journal, Vol. 5 No. 4, pp. 187-97.
APPENDIX
Figure 5: The stock flow diagram of an apparel supply chain using Vensim software.
Retailer
Distributor
Manufactu
r
e
r
ICORES2013-InternationalConferenceonOperationsResearchandEnterpriseSystems
346