Integrating System Dynamics and Enterprise Modeling to Address
Dynamic and Structural C omplexities of C hoice Situations
Arash Golnam', Ann Van Ackere’, Alain Wegmann!
'Ecole Polytechnique Fédérale de Lausanne
School of Computer and Communication Sciences (I&C)
Systemic Modeling Laboratory (LAMS)
Station 14, CH-1015 Lausanne, Switzerland
Tel: {+41 21 693 67 94, +41 21 693 4381}
Email {arash.golnam, alain.wegmann}@epfl.ch
*Université de Lausanne
Faculté des Hautes Etudes Commerciales (HEC)
Department of Operations (DO)
Quartier UNIL-Dorigny, Batiment Internef, 1015 Lausanne
Tel: +41 21 692 34 54
Ann. VanAckere@unil.ch
Abstract
The ability of enterprise managers to come up with the decisions leading to the best outcome for
the enterprise is hampered by their cognitive limits in understanding and addressing the dynamic
and structural complexities residing in choice situations. Dynamic complexities deal with the
behavior of the enterprise and its environment over time. Structural complexities, on the other
hand, arise from the number of the departments within the enterprise, their interactions and the
interactions between the enterprise and the entities across its boundary. The policy aiding
methods developed to assist managers in the analysis of choice scenarios address these two types
of complexities in separation. In this paper adopting a holistic approach, we integrate System
Dynamics (SD) - a method for understanding the behavior of systems over time - and Systemic
Enterprise Architecture Methodology (SEAM) - a modeling method that provides insights into
how an enterprise and its interactions with other entities are structured. Integrating SD and
SEAM, we present an approach to modeling, analysis and simulation of choice scenarios aiming
at reducing the dynamic and structural complexities involved in the decision making process. We
illustrate the applicability of our approach by applying it to an example of a choice situation in a
manufacturing company.
Keywords: Dynamic and structural complexity, System dynamics, Enterprise modeling
1. Introduction
In many cases, managers in enterprises are faced with choice situations where a decision should
be made under high levels of uncertainty. In such cases, managers’ ability to come up with the
decisions leading to the best outcome for the enterprise is hampered by their cognitive limits in
understanding and addressing the structural and dynamic complexities of the enterprise. To
understand the dynamics and structural complexities we should adopt a systemic view of the
enterprise. An enterprise as a system is composed of sub-systems and is embedded in a hierarchy
of larger systems such as segments, markets, industry and etc. Structural complexities arise from
the number of the subsystems of an enterprise as well as, the interactions within the internal
systems and between the internal and external systems across the boundary of an enterprise. The
dynamic complexity, on the other hand, deals with the behavior of the enterprise over time. In
some cases a change in the state of an enterprise system element leads to a chain of reactions by
other system elements and may take years to fully play out its effect due to the existence of
causal ambiguities and delays (Sterman 2000) . There is a significant need for approaches that
help managers gain a better understanding of and address such inherent complexities the and
thereby come up with decisions that can help the enterprise achieve its strategic objectives.
In this paper we propose an approach to address the challenges arising from decision making in
uncertainty. Our approach consists of a process that integrates an enterprise modeling technique
and system dynamics methodology (Forrester and Wright 1961). Figure 1 illustrates an overall
view of our proposed approach.
The process consists of six steps. In the first step, we conceptualize the as-is architecture and
develop the enterprise models. Then, the stock and flow diagram is built. We show how an
enterprise model of the as-is architecture can help us in building and understanding the stock and
flow diagram. Next, we develop scenarios that can capture the topic of interest pertinent to the
choice situation. Scenarios are then simulated using system dynamics simulation of the stock and
flow diagram.
In the second next step, we check whether the simulation of the scenarios has led us to the
identification of the problem residing in the as-is architecture. If a problem is identified we
proceed to the third step, otherwise we are facing one of the following situations: 1) the current
model (i.e. as-is architecture, the stock and flow and the scenarios) does not capture the existing
problem. In this case the model needs to be re-conceptualized we call this a re-scoping. Perhaps
we need to go to a higher or lower level of abstraction. 2) no problem exists and the model
reflects the right abstraction of reality. In this case, the process is terminated.
The third step comprises of two activities; 1) solution scanning that involves generating all the
alternatives that can serve as a solution to the identified existing problem, and 2) selection of a
solution that has the potential to address the identified problem. Similar to step 2, if a solution
can not be found we probably need to re-scope the model.
1.1 As-is Architecture Conceptualization
1.2 Stock and Flow Development
1.3 Scenario Development
1.4 Simulation
No No
~~ (model
does not
represent
problem)
. Any proble!
to address _
—»| — Rescope
| Yes
v
3.1 Alternatives of Potential
Solution Generation
3.2 Potential Solution Selection
No
Solution
\ Identified?
5.1 To-be Architecture Design
5.2 Stock and Flow Modification
5.3 Simulation
|
“6. Potential
Solution
falidated ?
End
No x
Figure 1: The six-step process
In the fifth step, we design a to-be architecture based on the potential solution identified, modify
the stock and flow diagram and re-run the simulation. Finally, we check whether or not the to-be
architecture is validated (i.e. the problems in the as-is architecture are tackled). If so, the process
is terminated, otherwise we need to re-scope the model.
The rest of the paper is structured as follows. In Section 2, we present a comparative analysis of
enterprise modeling and system dynamics drawing implications on how the two methods can
complement one another. In order to illustrate the applicability of our proposed approach, in
Section 3, we apply it to an example of a choice situation in a manufacturing company. In this
example the manufacturing company needs to make a tradeoff between price and delivery lead
time in responding to a request for quotation (RFQ). Finally, in Section 4 we present our future
work and the conclusion.
2. System Dynamics and Enterprise Modeling, the Perfect Integration?
In this section we compare system dynamics and enterprise modeling to draw inferences on how
the two approaches can complement one another. We also point out a number of implications on
the integration of the two methods. Next, we briefly introduce Systemic Enterprise Architecture
Methodology (SEAM). SEAM is the enterprise modeling method applied in our proposed
approach in this paper.
2.1 System Dynamics versus Enterprise Modeling
System Dynamics (SD) is a methodology for understanding the behavior of complex systems
over time. It provides fundamental contributions to framing, understanding, and discussing
complex issues and problems. System dynamics originally developed by Jay Forrester’s work at
MIT in the 1950s, is centered around modeling and simulating complex systems through
systemic representation of the system in terms of stocks, flows, and feedback loops. SD methods
provide “essential insight into situations of dynamic complexity,” especially when experimenting
the real systems is impossible or not feasible (Sterman 2000).
SD provide significant insights into the behavior of the system over time but does not provide
any implication on how the system elements should be reconfigured to yield a desired behavior.
SD, in other words, captures the “what” of the dynamic behavior and does not address “how” the
behavior of the system can be modified.
System dynamics Enterprise modeling
Allows interpretation and prediction of the | Provides a static image of the system elements
behavior of a system over time (i.e. | and the details of their interactions at a point of
prescriptive in nature). time. (i.e. descriptive in nature).
Focuses on modeling an issue or a problem. | Focuses on building models of the enterprise
“the system dynamics process starts from a for the purpose of documenting,
. communicating and sharing an understanding
problem to be solved” (Forrester and Wright
of an enterprise (Vernadat 1996).
1961). re )
Provides a limited set of graphical notations | Provides a set of constructs and graphical
(i.e. stocks, flows and information objects)
notations to capture systems, information and
material flows, processes, and etc.
A model of reality can be built in different
ways by different modelers.
Reference models and model building blocks
as well as the graphical notations lead to a
shared understanding of the enterprise model
developed.
Suitable for scenario and policy analysis.
Suitable for system design,
reengineering and enterprise integration.
process
Aims at developing models to assist the
understanding of a behavior of system at an
aggregate level and thereby helps reduce and
address the dynamic complexity.
Aims at developing models to assist the
understanding of the system elements and their
interactions at different levels of abstraction
and thereby helps reduce and address the
structural complexity
Table 1: Comparison of system dynamics and enterprise modeling
Enterprise modeling, on the other hand, aims at providing enterprise models that lead to a better
understanding of how an enterprise functions. Enterprise modeling techniques provide a static
image of a state of the system. Enterprise models capture a detailed representation of system in
terms of the system elements and their interactions to provide a means of understanding or
communicating of the enterprise. The modeler can decompose the enterprise to a consistent set
of modules and hence gain a better understanding of and manage the system complexity.
(Vernadat 2002) In Table 1, we compare system dynamics and enterprise modeling.
2.2 Implications on the Integration
System dynamics and enterprise modeling complement and supplement one another from a
variety of perspectives.
A major problem of the managers in enterprises concerns their bounded cognitive ability to
assess the impact of their various decisions over time due to the existence of the dynamic
complexities. Scenario analysis and development methods are therefore needed to evaluate
different scenarios arising from the choices available to the managers. System dynamics provides
a rigorous basis for modeling the problems in which reside the choice situations as well as
developing and comparing various choice scenarios to one another.
However, an enterprise may have to reconfigure itself or eliminate the redundant activities in
order to be able to achieve the outcomes associated with a desired scenario. By modeling various
aspects of an enterprise such as information flows, functions, processes and etc. at different
abstraction levels, enterprise modeling provides a blue print for an enterprise to migrate from an
as-is to a to-be situation yielding the desired outcomes.
Moreover, an enterprise model can serve as a good starting point for building a stock and flow
diagram. Enterprise models provide a rigorous and consistent abstraction of reality based on
which a stock and flow diagram can be developed. An enterprise model accompanying a SD
stock and flow diagram can help communicate the context and the focus of the problem and
make the model more understandable.
2.3 Systemic Enterprise Architecture Methodology (SEAM)
In this paper we apply Systemic Enterprise Architecture Methodology (SEAM) (Wegmann
2003). The foundations of SEAM are in General Systems Thinking (GST) (Weinberg 2001) and
in RM-ODP (ISO Standard 1995). GST is the study of principles that are applicable to any kind
of system (e.g. business system or IT system). RM-ODP is a software engineering ISO standard
that provides solid definitions for the SEAM concepts (e.g. process, state, property). SEAM is
rigorously defined based on these systemic and software engineering concepts. SEAM federates
multiple modeling techniques (such as discrete behavior, goals or quantitative models). SEAM
has been applied for teaching (Wegmann et al. 2007) and consulting since 2001.
SEAM augments the enterprise modeling methods by modeling the environment in which the
enterprise is embedded and hence creates the possibility and supports modeling of the business
strategies. In SEAM, an enterprise is viewed as a nested hierarchy of systems. At a micro level
we model the enterprise, the various departments inside the enterprise and the processes within
each department. At a macro level we model and analyze how companies working as a value
network can cooperate to achieve commercial objectives.
SEAM provides a consistent set of modeling principles and constructs to model an enterprise at
different abstraction levels. In SEAM the systems (i.e. value segments, value networks,
enterprises and the departments) can be represented as wholes (black boxes) abstracting the
components or composites (white boxes) showing the system components. Whole and composite
representation helps the modeler simplify the conceptualization of the structural complexities
inherent in the business environments.
3. Applying the Proposed A pproach to the Example of BE Co.
To illustrate the applicability of our proposed approach and provide insights into how enterprise
modeling can be integrated with system dynamics we apply the 6-step approach as presented in
Section1 to the example of a choice situation in a manufacturing company called BE Co. This
example is a fictitious case inspired by a real company.
BE Co. (Best Engine Company) is specialized in designing, manufacturing and servicing diesel
engines for light aircraft. NewPlane SA which is one of the most promising plane manufacturer
companies in the aeronautic business has decided to buy 240 units of aircraft diesel engines over
a two year period of time. NewPlane SA has sent a request for quotation (RFQ) to BE Co. and its
all time competitor QEng Co. (Quality Engine Company). BE Co. has had a number of
experiences taking part in the tenders organized by NewPlane SA and it also has a good
knowledge of the market in particular its competitor QEng Co. On this basis, BE Co. holds the
following assumptions:
« NewPlane SA always chooses to source its demand from at least two suppliers.
e Diesel engines for light aircrafts are also supplied by QEng Co. Hence, the demand for
240 units of engines is going to be split between the two companies.
e As there is not considerable quality difference between the engines it manufactures and
the ones manufactured by QEng Co., NewPlane SA makes its choice on the basis of
delivery lead-time and price.
e QEng Co.’s price can range from 120 to 140 K USD for each unit of engine and the
delivery lead-time is between 3 to 5 months.
BE Co. has to procure the engine parts from the suppliers. For the sake of simplicity we assume
that only two types of parts are required for engine manufacturing; a kit and a mount. It has to
figure out the right tradeoff between the lead time and price to get the maximum number of
orders at a competitive price and lead time while ensuring profitability.
3.1 Step 1
3.1.1 As-is Architecture Conceptualization
In this step we apply SEAM to conceptualize the as-is architecture of BE Co. Before explaining
the as-is architecture, we briefly present a number of modeling principles used to develop the
conceptualization of the as-is architecture. In SEAM block arrows represent systems. When
modeling a business context we represent a company and its value network as systems. A value
network is a group of companies that collaborate and pursue a common commercial objective.
Systems can be represented as a whole (black box view of a system) or as a composite (white
box view of a system). Modeling a system as a whole, the system components and their
interactions are abstracted. Instead, the system’s emergent actions and properties that
conceptualize the overall behavior of the system are observed. On the other hand, when a system
is modeled as a composite, its structure, component systems and their interactions are viewed.
Figure 2 is a conceptualization of the as-is architecture of BE Co. value network as composite.
Modeling a value network as a composite, we capture the interactions between the companies
inside the value network to gain a better understanding of how the companies cooperate to
deliver a product or service to the customer. In Figure 2, inside the value network we have
modeled BE Co., Mount Supplier X and Kit Supplier Y as wholes, thereby showing their
emergent processes and abstracting the details of the processes and the departments inside each
company. In SEAM, a process is represented by an ellipse. As shown in Figure 2, BE Co. has the
core process of “Engine design, manufacturing and sales” which triggers the “Order setting”
process that recurs for kits and mounts. Since, BE Co. is within the scope of our analysis we
7
have modeled the processes of the other entities within BE Co. to the extent that fits our purpose
of analysis.
Modeling a system as a whole we can also represent its emergent properties. Emergent properties
can be material and/or information that are output and/input of processes. Inspired by the work of
(Dori 2002), material or information properties are respectively denoted by double line and
single line rectangles. In Figure 2, an “Engine Order” is an information property that is inputted
to the “Engine design, manufacturing and sales” process within Be Co. This process, in turn,
outputs the “Product BOM” as an information property which triggers the “Order Setting”
process for mounts and kits. Similarly the orders placed by BE Co. (i.e. information property)
serve as the inputs that trigger the “Manufacturing” process within Supplier X and Supplier Y
company which output “Kit” and “Mount” as material properties.
BE Co. VN Ic] Eng Co. tw
Supplier X [W]
&Sales dl
Mount Order Z <
p
/Nalue Network
Performance
BE Co. [w] NewPlane SA [w]
i
Pane Des
‘ote varus
* eral
Engine Design,
Product BOM
Manufacturing & Sales
{ T
Order Setting for 4”
Kits
Kit Manufacturing
/Nalue Network
Supplier Y fw]
Kit Order
LEGEND
= System ---> Information Flow
© Process —
[1] Property —>[E J] Material Property
a“)
[w] Systemas a whole
Material Flow
[c] Systemas a Composite
eo Multi-Party Relationship Information Property
Figure 2: BE Co. Value Network as a composite (as-is architecture)
Other sections of the model can be interpreted the same way. The inflow and outflow of material
and information properties are captured by dashed and dotted lines respectively. The dotted line
from “Engine order” information property and “Kit” and “Mount” material properties to the
“Order setting” processes imply that the “Order setting process” operates on the basis of the
information inputs from these entities.
Processes also have emergent properties. In the case of the companies within BE Co. VN such
emergent properties are the delivery lead time and price. We use a diamond to represent the
multi-party relationship (MPR) capturing the overall performance of the companies within the
value network. MPRs can also represent the company level performance. As it can be seen in the
model, with the current lineup of suppliers, BE Co. can deliver the engine at the price of 140 ina
4 month period.
When modeling a value networks we also represent the customer and the competitor companies.
In Figure 2, we can see the flow of orders and engines between NewPlanes SA and BE Co. We
can also see the overall performance of the QEng. Co. (based on the assumptions presented at the
beginning of this section) value network as the competitor company.
3.1.2 Stock and Flow Development
At this step we develop the stock and flow diagram. Figure 3 exhibits the stock and flow diagram
of BE Co. supply chain. As mentioned earlier, the enterprise model can provide useful insights
into developing the SD model. More precisely, the processes in the enterprise models correspond
to the rates in the stock. In addition, the information and material properties can be mapped onto
the stocks. Further, the flow structure is almost similar in both stock and flow diagram and the
enterprise model. As an instance to clarify this similarity, “Kit order rate” in the stock and flow
diagram corresponds to the “order setting for kits” process in the enterprise model and the “Kits
ordered by BE Co.” stock can be mapped onto the “Kit Order” information property in the
enterprise model. The mapping between “Kit manufacturing and delivery” rate in the stock and
flow diagram and “Kit manufacturing and sales” process in the enterprise model can be
interpreted the same way. As it can be seen in Figure 3, we have structured the stock and flow
diagram in such way that the commonalities can be easily identified.
Figure 3: BE Co. Value Network as a composite - stock and flow diagram capturing the
production aspects
We use the stock and flow diagram in Figure 4 to model the cash inflow and outflow in BE Co.
In this model, OpEx is the operating expenses. The operating expenses for the production of each
engine is 50 K USD. We have also assumed that the engine manufacturer pays the suppliers
when its engines are sold to NewPlanes SA.
10
Price per mount
(KS)
a Cost of mounts BE Co. tou
Cash oulow rate Cash iow rate
\
PH cost oki ioe :
Total Opes
—
| “a
Ope per Engine
Figure 4: BE Co Value Network as a composite - stock and flow diagram capturing the financial
aspects
3.1.3 Scenario Development
As mentioned earlier, BE Co. knows that the orders it can win from NewPlanes SA is a function
of its lead time and price relative to the lead time and the price of QEng Co. Hence, to develop
the scenarios inspired by the work of (Kim 1998), a nonlinear demand split function that can
estimate order quantities for BE Co. and QEng Co. is used. By being nonlinear, the impact of the
price and lead time factors of a company on the orders won becomes disproportionate. Thereby,
a slightly better lead time or price can lead to a considerable increase proportion of the demand.
The non-linear function we applied in our model is a power function. We have formulated the
order split ratio for BE Co. as follows:
(QEng. Co. Leadtime * QEng. Co. Price)?
(BE Co. Lead — time * BE Co. Price)? + (QEng. Co. Leadtime * QEng. Co. Price)?
BE Co.'s Order Ratio =
As mentioned earlier QEng Co.’s price ranges between 120- 140 K USD with the delivery lead
time of 3 to 5 months for each unit of engine. On this basis we have developed two scenarios for
BE Co. with the as-is price of 140 K USD and the delivery lead time of 5 months. Table 2
summarizes these scenarios.
11
QEng Co. | Lead time | Price (K
Scenario (Months) USD)
Best case scenario 5 120
Worst case scenario 3 140
Table 2: The two scenarios
In Table 2, best case scenario means that if QEng Co.’s lead-time and price is 5 months and
120K USD respectively, the best case has occurred for BE Co. In that case, BE Co. can get a
higher demand split ratio. It should be noted that since there is a tradeoff between price and lead-
time it is not possible to decrease the lead time and price simultaneously.
3.1.4 Simulation
We have simulated the cash level at BE Co. over a 2 years (24 month) period of time for the two
scenarios. The results of the simulation are illustrated in Figure 5.
BE Co. Cash As-Is Situation
2,000
1.500 Actual Cash Level BCS: 1800
. K$
4000 Actual Cash Level WCS: 1200
— K$
500
0
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Month)
Cash : Best case scenario
Cash : Worst case scenario
Expected cash level (3000 K$)
is not reached > problem
Figure 5: BE Co.’s cash level over a two year period of time for the as-is situation
3.2 Step 2 Any Problem to Address ?
With the as-is architecture, and in the best case scenario (BCS) BE Co. can only get around 80
orders ( around 42 % of the total orders ) and reach 1,800 K USD level of cash which is far
below expected cash level of 3000 K$. The expected cash level if calculated considering the rest
of the expenditures BE Co. has such as CapEx (Capital Expenditures) and etc. Hence, it is
apparent that a problem has been identified.
12
3.3 Step 3
3.3.1 Alternatives of Potential Solution Generation
In Step 2 BE Co. realized that with its current lead time and price it can not win enough engine
orders to reach its expected cash level. Hence, it has to reduce its lead time or price in order to be
able to capture a larger proportion of NewPlanes SA orders and thereby raise its level of cash.
For BE Co. it is evident that reducing the price is out of question as the price it is planning to
quite is very competitive. Hence, based on the model so, BE Co. can pursue two different
strategies that can serve as solution alternatives.
Solution Alternatives
1 | Reduce the supply lead-time for kits (currently 2 months) — no need to change part lead-
time as not in critical path
2 | Reduce its manufacturing cycle time (currently 3 months)
Table 3: Solution alternatives for BE Co.
Since BE Co. has been engaged in a long time partnership with its suppliers and has in fact
invested in training programs in its supplier companies to enhance the quality of the parts
supplied, changing suppliers will not be feasible at all. Hence, it has to reengineer its internal
process to improve its as-is situation.
3.4 Re-scope
As the current model scope the (i.e. the as-is architecture, the stock and flow and the scenarios)
does not capture the details of the internal processes at BE Co., we need to re-scope the model in
order to be able to select a solution based on which the to-be architecture can be designed. Hence
we move back to Step 1.
To gain an insight into the internal processes within BE Co. and to find out whether or not any
improvement can be made we conceptualize BE Co. as a composite. Fig. 6 presents a SEAM
model of BE Co. as a composite. The composite view of BE Co. provides useful insights into the
component processes of “Engine design, manufacturing and sales” process which is the emergent
process in the BE. Co as whole. As we can see the core emergent process of “Engine design,
manufacturing and sales” is now decomposed to six sub-processes (i.e. Order Processing,
Engineering, Procurement, Manufacturing, Testing and Accreditation and Shipment) carried out
in four departments (Sales and Shipment, Engineering and Manufacturing, Procurement and
Inventories, Homologation and Type Approval).
As this change in the as-is architecture does not precipitate any change in the structure of the
stock and flow diagram, the previous stock and flow diagram can remain intact and no
13
modification is required. In Step 2, we have the identified problem of long lead-time and high
price. In step 3, we generate the solution alternatives based on the re-scoped model.
BE Co. [c]
Procurement and Inventories [w] Sales and Shipment [w]
Procurement
1
Engineering and Manufacturing [w]
Order Processing
1/8 Month
Tech:
Spee.
Shipment
Company
Homologation and TA [w]
A
Manufacturing
Figure 6: BE Co. as a composite (to-be architecture)
Table 4 summarizes the solution alternatives (2.1 to 2.4) that have the potential of addressing the
problem no. 2 identified in step 2.
Solution Alternatives
2.1 | Reducing the 1 Month manufacturing lead time
2.2 | Reducing the 1/4 month Sales and Shipment lead time
2.3__| Reducing the 1/4 month Procurement lead time
2.4 | Reducing the 3/2 months Testing and Homologation lead time
Table 4: Solution alternatives for BE Co. — Model re-scoped
After scanning all possible solutions to the problem, we now evaluate the solution alternatives in
order to come up with the solution that can meet an initial feasibility criteria.
Solution 2.1: Reducing the 1 Month Engineering and Manufacturing lead time:
BE Co. has recently invested in manufacturing equipment such as automated guided vehicles
(AGV) that led to a major reduction in its manufacturing cycle time. An effort to further decrease
the manufacturing cycle time required an additional capital expenditure of 1,300 K USD to
14
increase the automation level of the production line by employing a flexible manufacturing
system (FMS). This will also require training courses for the engineering and manufacturing
personnel. Altogether, BE Co.’s management is not willing to move towards this solution due to
the high level of investment and the time required to move up the learning curve.
Solution 2.2: Reducing the 1/4 month Sales and Shipment lead time:
Order processing involves a careful elicitation of the customer requirements and sending the
technical specification requirements to the engineering department. The shipment process, on the
other hand, involves the packaging and preparation of the required paper work for the product to
be shipped to the customer. The one week deadline has been achieved after years of experience
in BE Co. and can not be squeezed any further.
Solution 2.3: Reducing the 1/4 month Procurement lead time:
The procurement process which is composed of sending RFQs to the suppliers, receiving
feedbacks and placing orders takes almost a week in BE Co. It normally takes the suppliers a
couple of working days to respond to the RFQs sent by BE Co. their quotations should then be
analyzed by BE Co. and orders are then placed accordingly. Since a part of this one week lead
time is geared to the response time by the suppliers it is not possible for BE Co. to decrease it.
Solution2.4: Reducing the 3/2 months Testing and Homologation lead time
Regulation plays an important role in the aeronautics business. Aircraft engine manufactures
need to conduct a wide variety of tests on their engines to ensure that their engines conform to
the safety standards and emission level requirements. The test results need to be accredited by a
third party so that the engines can be homologated and type approved. Currently BE Co. carries
out all the tests and sends the test results to an accreditor. BE Co. can reduce the testing time by
one month if it outsources the testing as well as the homologation activities to a third party tester
and accreditor.
3.5 Step 4 Potential Solution Identified?
As outsourcing the testing and homologation was selected as a potential we move on to the next
step.
3.6 Step 5
3.6.1 To-be Architecture Design
As outsourcing the testing and accreditation was selected by BE Co. as a solution that can
contribute to shortening the lead-time, the to-be architecture is designed accordingly. Figure 7
depicts the to-be architecture of BE Co.’s value network. As it can be seen Tester and Accreditor
15.
Co. is the new entity that is introduced in this model. BE Co. sends the manufactured engines to
Tester and Accreditor Co. and the required homologation documentation is sent back to Be Co.
in about two weeks. As it can be seen in the MPR, This re-configuration leads to a month of
reduction in lead time and an increase of 15 K$ in the price.
3.6.2 Stock and Flow Modification
Be Co.’s price and lead time in the to-be situation will respectively be 155 K USD and 4 Months.
We modify the stock and flow accordingly.
3.6.3 Simulation
We run the simulation of the stock and flow diagram that captures the to-be architecture to see
whether the results lead to any improvement in terms of bridging the gap between the actual and
the expected cash level.
BE Co. VN [c]
QEng Co. fw)
Supplier X (Wd
Tester and Accreditor Co. [wl
fount Manufacturing
Sales
Month | [SKS
Testing and TA
‘Value Network
BE Co. [vd
NewPlane SA [w]
(Order Setting for
Mounts
Plane Design,
Homologatian Manufacturing &
Doc. Sales
4 }
+
Engine Design,
Manufacturing & Sales
Engine Order
Product BOM
(Order Setting for
kits
Kit Manufacturing
‘Supplier ¥ [1
2 =
Figure 7: BE Co- Value Network (to-be architecture)
3.7 Potential Solution Validated?
Figure 8 compares the scenarios in Table 2 for the to-be architecture. As it can be noticed, the to-
be scenario yields satisfactory results in terms of the cash level. Comparing the best case
16
scenarios of the as-is and the to-be architectures , BE Co. can almost capture 50% of the orders
placed by NewPlanes SA and double its stock of cash.
BE Co. Cash To-Be
4,000
Actual Cash Level BCS:
3,000 3800 K$
S 2,000 Actual Cash Level
WCS: 2500 K$
1,000
0
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Month)
Cash : Best case scenario to-be
Cash : Worst scenario to-be
Expected Cash Level
(3000 K$) achieved.
Figure 8: BE Co.’s cash level over a two year period of time for the to-be architecture
4) Conclusion and Future Work
In this paper we presented an approach for dealing with structural and dynamic complexities in
choice situations. Our approach consists of six-step process in which we migrate from an as-is to
a to-be architecture that incorporates a selected solution to an existing problem in the as-is
architecture. Adopting the view of an enterprise as a system composed of subsystems (such as
departments) which is in turn nested in a hierarchy of a larger systems such as value networks, a
multi-scope analysis should be conducted in order to address the various structural complexities
that can arise from the interactions between the entities within the enterprise as well as the
interactions of the enterprise with the entities outside its boundary. Hence, identifying the
problem in the as-is architecture and designing a to-be architecture that can address the problem
an iterative re-scoping of the conceptualization should be done. Systemic Enterprise Architecture
Methodology (SEAM) as an enterprise modeling method equips us with the ability of
conceptualizing and designing the architectures at different levels of abstractions. SD on the
other hand created the possibility of problem identification and solution validation through
simulation of scenarios.
17
Our future work focuses on applying our approach to a prospective business case to check its
applicability in real business settings. We will also examine how a better mapping between our
enterprise modeling approach and SD can be made and how the two techniques can complement
one another. Exploring the relationship between the modeling approaches can lead to significant
insights into how the two methods can be integrated.
5) References
Dori, D. 2002. Object-process methodology: a holistic systems paradigm. Springer Verlag.
Forrester, J. W, and J. Wright. 1961. Industrial dynamics. MIT press Cambridge, MA.
ISO Standard. 1995. ISO/IEC 10746-1, 2, 3, 4| ITU-T Recommendation X. 901, X. 902, X. 903,
X. 904.“. ITU. Google Scholar.
Kim, B. 1998. A System Dynamics Approach to Decision Making on Outsourcing from
Suppliers with Different Innovation Capabilities. In Conference Proceedings.
Sanchez, Ron, and Aimie Heene. 2003. The New Strategic Management: Organization,
Competition, and Competence. Wiley, July 21.
Sterman, John D. 2000. Business Dynamics: Systems Thinking and Modeling for a Complex
World. McGraw Hill Higher Education, December 1.
Vernadat, F. 1996. Enterprise Modeling and Integration: Principles and Applications. Ist ed.
Springer, January 15.
Vernadat, F. B. 2002. Enterprise modeling and integration (EMI): current status and research
perspectives. Annual Reviews in Control 26, no. 1: 15-25.
Wegmann, A. 2003. Systemic Enterprise Architecture Methodology (SEAM). In Published at
the International Conference on Enterprise Information Systems 2003 , 483-490.
Wegmann, A., G. Regev, J. D. De la Cruz, L. S. Lé, and I. Rychkova. 2007. Teaching Enterprise
and Service-Oriented Architecture in Practice. In Proceedings of the Workshop on Trends
in Enterprise Architecture Research (TEAR 2007).
Weinberg, G. M. 2001. An introduction to general systems thinking (silver anniversary ed.).
Dorset House Publishing Co., Inc. New York, NY, USA.
18
