Attempt to Integrate System Dynamics and UML in
Business Process Modeling
Liang-Cheng Chang
Department of Information Management
National Sun Yat-Sen University, Kaohsiung 804, Taiwan.
886-7-5252000 ext. 4760
d9142809@student nsysu.edu.tw
Yi-Ming Tu
Department of Information Management
National Sun Yat-Sen University, Kaohsiung 804, Taiwan.
886-7-5252000 ext. 4717
ymtu@mis.nsysu.edu.tw
Abstract
The integration of information systems and business process will affect competitive
advantages of firms. In order to develop information system, modeling of business
process is a fundamental work of system analysis and design. System dynamics is
useful to solve non-linear, complex, time delay and feedback problems of business
processes. However it still belongs to a special field of modeling language because it
cat be integrated well with information systems in organizations. The purpose of
this paper is to integrate system dynamics with UML and thus they can be developed
synchronously during information systems implementation in enterprise. For this
reason, integrated development process and system architecture with system
dynamics and UML have also been proposed in this paper
Keyword: System Dynamics, UML, Object-Orientation, System Analysis and
Design
L. Introduction
Every company has a number of business processes which is some kinds of
activities within the organization where people or machines work together to achieve
desired outcome or their business objectives (Ould, 1995). For activities in business
processes, Hammer (1996) classified activities into three types, Value-adding work,
Non-value-adding work, and Waste. Porter (1985) divided activities into support
activities and primary activities. Support activities include firm infrastructure,
human resource management, technology development, procurement and Primary
activities include inbound logistics, operation, outbound logistics, marketing and
sales, and service. Porter (1985) argued that firms can raise competitive advantages
from primary value added activities. Therefore, the design of business process would
affect competitive advantages of firms.
Harmer (1990) proposed a concept of Business Process Reengineering that is the
radical redesign of business processes to cut non-value-adding work and waste. The
main concept of reengineering is to integrate business process with information
technologies. Thus traditional paperintensive process must be broken, and be
redesigned in the center of information technologies. In order to make business
process to adapt to information technologies, many business process modeling
methodologies have been proposed.
Rumbaugh (1999) pointed out that model is a representation in a certain medium
of something in the same or another medium and model captures the important
aspects of the thing being modeled from a certain point of view and simplifies or
omits the rest. A model of a software system is made in a modeling language and it
has both semantics and notation and can take various forms that include both
pictures and text.
Recently, some technologies to model real world and to develop software
systems have been presented, Object-Orientation, Function-Orientation,
State-Orientation and Formal (Peters, 2000). Object-Orientation is the most popular
method which could make software product to be reusable, portable and
interoperability (Schach, 2002). For enterprises, if they model business process by
object method, they can adjust their system more flexibility.
UML (Unified Modeling Language) is the standard object-oriented modeling
language proposed by Booch, Jacobson and Rambaugh. During the software
development process, software engineers can communicate to each other through
UML which includes 10 diagrams by static and dynamic views. In this paper, we
apply UML to model business process because it has been the standard modeling
language of OMG (Object Management Group).
Besides previous methodologies, System dynamics is another methodology to
model business process. Senger (1990) identified business process as a circle of
causality that describes a feedback loop of cause and effect, and it could be depicted
by system perspective. Sequentially, Sterman (2000) argued that system dynamics can
be applied to model business process and to solve complex problems included
Constantly Changing, Tightly Couple, Govemed by Feedback, Non-Linear, History
2
Dependent, Self-Organizing and Counter intuitive. As we known, most practical
business processes have above characteristics. Therefore, system dynamics seems to
be more powerful than normal software modeling methodologies especially in
dynamic and non-linear processes and most of business problems all have these
features.
In spite of system dynamics has a lot of advantages to become a model language
of business information systems, it has still been confined in a specific filed and
doesn’t belong to a popular software modeling language. Tu (2003) pointed out the
limitation of traditional simulation software of system dynamics. Firstly, the isolated
interface of traditional software doesn’t support the demand in decision of the enterprise
today to update current data from database. Secondly, when traditional software simulates
each policy or scenario, it must reset the data and execute the simulation process every
time. Besides, most traditional software products were programs of single computer and
they couldn’t be integrated with business information systems through networks. If
companies want to apply system dynamics to be a powerful tool to solve their problems,
they must integrate system dynamics with their information systems. It would be better to
develop information systems and system dynamics synchronously in the beginning.
Based on previous reasons, the purpose of this paper is attempt to integrate system
dynamics with UML to model business process and also to propose an integrated
information system development process by system dynamics and UML. The second
paragraph of this paper is to compare components of object-oriented methodology and
system dynamics. The third paragraph is to transform causal diagram and stock-flow
diagram to UML diagram by the case of market-growth in order to show that stock flow
diagram and UML diagrams can be exchange each other. Fourth paragraph is to propose
an integrated development process of system dynamics and UML. Conclusion and
suggestion will be presented in last paragraph.
2. Comparison of Object-Orientation and System Dynamics
The concept of object-orientation is a revolution of software development.
Object-Oriented Analysis and Design (OOAD) makes software development time
shorten and system maintain easily. By the definition of OMG object is a
combination of a state and a set of methods that explicitly embodies an abstraction
characterized by the behavior of the relevant requests (OMG, 1991).
A main feature of object-orientation is to consider both data and actions to be
equally important and to combine data and action in Class that is a template for
several objects and presents intemal structure of object (Jacobson, 1992;
3
Sommerville, 2004). A class is composed by Attribute and Behavior (Jacobson,
1992). Data in class are called Attributes which mean characteristics in order to
describe state of class. By the perspective of information system, attributes are data
or information gathered from database or inputted from users. Actions in class are
called Behavior or Operation that means the action or method to stimulate or to
change the state of class. Here, we apply the item “operation” in this paper By
dynamic view of class, attributes may be affected by operation.
Components of system dynamics are Sector, Level, Rate and Auxiliary. Level or
Stock characterizes the state of the system and provides the basis for actions. It also
provides system with inertia and memory (Sterman, 2000). Rate or Flow is the
change of level with time. By the definition, rate is the action to change the state of
the level with time.
Class can be regarded as the sub-system of information systems from the
definition of class in object-orientation and sector can also be regarded as
sub-system in system dynamics. Therefore, we can find that class and sector have
similar characteristic which is the sub-system of the whole system. They can also be
denoted by mathematic description of Set { }.
The characteristics of Level and Attribute are also similar and they can both
present the state of system. By the definition, level and attribute can be denoted by a
variable Y which means a state of system. Besides, auxiliary can also express the
property of system or it can be set as an initial condition or other variables affected
by environment. So an auxiliary can also be regarded as an attribute because they
both present data or information. Auxiliaries can be denoted as a constant C, other
variable X, function of level Y, function of other variable X or function of time.
The characteristic of Rate and Operation is similar and it presents the action
which changes the state of system. By the definition, the mathematic symbol of Rate
or Operation is differentiation of level Y with time.
In summary, the comparison of features in system dynamics, object-orientation
and their mathematic description are summarized in Table 1.
Table 1 Comparison of system dynamics and object-orientation
aan Dynamics Objection —
Sector Class Set{ }
Level (Stock) Attribute Y
Rate (Flow) Operation (Behavior) s
Auxiliary Attribute C+ X + FY) + FX) + Fe)
3. Transformation of Stock Flow Diagramand UML Diagram
3.1 Introduction of SD diagrams and UML diagrams
The diagrams of system dynamics (SD) include casual diagram and stock flow
diagram. During the process of system dynamics, the problem of system is portrayed
by casual diagram and next step is to model and to simulate the problem by stock
flow diagram.
Unified Modeling Language (UML) is a standard language in order to model
software system and it was integrated by Rational from object-oriented modeling
tools proposed by Booch, Rambaugh and Jacobson. The purpose of UML is
specifying, constructing and documenting object-orientation system by visualizing
(Booch et al, 1999). There are 9 diagrams of UML and 5 perspectives of system
architecture presented in Table 2
Table 2 UML diagrams and system perspectives (Reed, 2002)
Perspective UML Diagram Application
Use Use Case Diagram Events of users, system.
analyzers, or system engineers
to use the system
Design Class Diagram, Object | To design the functions of
Diagram system or the service to users of
system
Process Sequencing Diagram, _| To present the synchronic
Collaboration Diagram, | sequence or process of system
State Diagram, Activity
Diagram
Deployment Component Diagram. To present the components or
files of system
Inplementation | Deployment Diagram _| To present network
infrastructure or hardware of
system implementation.
The connection of system dynamics diagrams and UML have been exhibited.
during past years. Tignor (2003) compared the features of activity diagram of UML
and stock flow diagram of system dynamics and then built a stock flow diagram
based on UML activity diagram. The main contribution of Tignor is to transform
activity diagram to stock flow diagram.
Sequentially, Tignor (2004) applied the case of British Telecom Intelligent
Network from UML diagrams to causal diagram and stock flow diagram. Tignor
also found the possibility of applying system dynamics to the problems of modeling
information systems.
Previous researches showed the possibility of transformation from UML
diagrams to SD diagrams. However, the study of transformation from SD diagrams
to UML diagrams is still scarce. The purpose of this paper is to apply a model of
system dynamics to transform from SD diagrams to UML diagram.
3.2 Simplified Market Growth Model
The part of the model “Market Growth” built by Forrester in 1975 would be
applied in this paper Original model included four causal loops, marketing effort,
market perception, production capacity utilization, and capital investment which
could be regarded as four business processes. In this paper, there are two loops
drown out from original model and be depicted as Sales and Orders. The casual
diagram of simplified market growth model is shown in Figure 1. In the loop of
sales, hiring salesman will increase number of salesman and more salesmen will
eam more orders booked. Then more orders will get more budgets of sales and more
budgets can let company hire more salesmen. We could find it is a positive loop for
sales loop. In another loop of orders, more orders booked will take more backlogs
and more backlogs will bring more delivery delay. There is a time delay between
delivery delay and sales effectiveness and a negative effect from delivery delay to
sales effectiveness. At last, better sales effectiveness will eam more orders booked.
We could find it is a negative loop for orders loop.
Orders Delay
Sal Booked.
neo Ys)
+ .
y + Delivery
[ a oe
es I i
Sales Expected. + + impending
Adjustment Number of by acldog __ ew
Salesman. -
Orders
(o married Complete
Salesman. Ss
Salary
Figure 1 Causal diagram of Simplified Market Growth Model
Then, the stock flow diagram is built in Figure 2. The loop “Sales” has been.
depicted as sector “Sales” and another loop “Orders” has been depicted as sector
“Orders”. In the sector of salesman, the number of salesman has been depicted as the
level of salesman, and hiring salesman has been depicted as the rate of salesman
hiring. Others items are auxiliaries. In the sector of orders, backlog has been
depicted as the level of backlog and delivery delay is another level. Order entered
and order complete are the rate of backlog and changes in DDR (Delivery Delay
Recognized) is the rate of delivery delay recognized. Auxiliaries in sector of orders
are also shown in Figure 2.
po) | ae ee os =
Figure 2 Stock- Flow diagram of Simplified Market Growth Model
The simulation result of this simplified market growth model is presented in
Figure 3. The quantities of Sales, Backlog, and Delivery delay recognized are
presented an S-shape growth, and sales effectiveness depresses and falls during the
time. Two loops of sales and orders could be regarded as business process and
processes of sales and orders affect each other and it becomes the system for an
archetype of limitation of growth proposed by Senge (1990). It’s hard to recognize
this kind of system complexity in business process by traditional information
systems, but system dynamics can. Therefore, implement system dynamics into
traditional information system would be inevitable task.
BB i ecg 2: Dekery Delay Recogrized 3 Saksnen
& §5000.00 >
2 500 = +
3 10000 Va |
45000.00 i 2
wr
By
3
i
Ne
N
8
onl 7
10.00 T
100 0.5 100.50 15025 200.00
Hl Graph (Utieg Weeks 1150 206583528
Figure 3 Simulation Result of Simplified Market Growth Model
33 Modeling Simplified Market Growth by UML
We will build UML diagrams based on SD diagram that is an opposite way to
previous researches. Firstly, by the definition of Table1, components of level, rate
and auxiliary in stock diagrams should be transformed to class, attribute and
operation in object-orientation concept. The results of SD diagram components
corresponding with Object-Orientation Components are shown in Table 3
Table 3 SD Components Corresponding with Object-Orientation Components
Components of Com of Component of Stodk-Flow Diagram in
Object-Orientation. i" Figure 2
Dynamics
Class Sector Sales Orders
SalesAdjustTime | Order Booklog
Sales
Expected Sal
% Effectiveness
Attribute Auxiliary Salesman Salary Time for DDR
Deivey Deday
Revenue from Sales “
Impending
Sales Budget Delivery Rate
Sal Dewey Delay
Attribute Leva Recognized
Backlog
Salesman Hiring Order Entered
Operation. Rate Order Complete
Change in DDR
Secondly, according to the content of Table 3, class diagram based on stock-flow
diagram of simplified market growth could be depicted as Figure 4. Two sectors of
stock-flow diagrams would be transformed to be two classes, sales and orders in
class diagrams and attributes and operations are also be shown as below.
Orders
Sales |+OrdersBacklog : float
“Sal Tink=10 |4+DeliveryDelayRecognition : float =2
ee ss 4Backlog : float =8000
|+SalesA djustTime : float =20 :
‘ -+TimeforDDR : float
-+ExpectedSales : float 5 ;
. = +DeliveryDelayImpending : float
+SalesmanSalary : float =2000 * float
-RevenuefromSales : float = 10 DelvayRay: ni
—— OrdersEntered) :
eee ) = Bloat +OrdersComplete() : float
+ChangeinDDR() : float
Figure 4 Class Diagram of Simplified Market Growth Model
Thirdly, based on stock-flow diagram in Figure 2 and class diagram in Figure 4,
sequencing diagram could be depicted as Figure 5. System users input data and get
result from system through user interface. The simplified market growth model can
be built as a marketing information system. According to Figure 2, rate of salesman
hiring is affected by sales adjust time and expected sales, so user input the data of
these attribute to system, and then get information of salesman from system. The
information of salesman would be transferred from class sales to class orders. After
9
user input related information into system, user can get information about backlog
and delivery delay recognized from information system.
During the process from stock flow diagram to sequencing, we found some
problems about transformation. First, stock flow diagram is synchronous, but
sequencing diagram is asynchronous. It is difficult to decide which attributes or
operations should be list first in sequencing based on stock flow diagram. Second,
sequencing diagram didn’t have the concept of time delay, but stock flow did. It's
difficult to describe time delay in sequencing diagram.
Userinterface Sales Orders
Users
[] Request 7 7 7
informetion 1 !
1 SaleAdjustTime ' '
f s '
' Salesman ' H
f ' '
' ExpectedSales | '
s+
' SalesHliring( ) 1 '
—/ '
1 Salesman ' sa t
je__ Samm lesman '
' 1 {
<4
' OrdersBacklog i
' Order Entered ) f
I
1 Backlog i
\
} Order Complete ) 1
\
t t
' Backlog ;
! Time for DDR 1
t fi 4
! Change in DDRO 7
j i +
i t *
le Delivery Delay Recognized !
' - -
' H H
' ' H
L ' ' '
'
Figure 5 Sequencing Diagram of Simplified Market Growth Model
Fourthly, based on sequencing diagram, activity diagram can be depicted as
Figure 6 which shows different activity states of system. It starts from hiring
salesman and this activity state will affect state of salesman. Then, state of salesman
will affect state of order booked which branches to budget and to order entered. On
state of backlog, the flow branches again to delivery rate and delivery delay
impending. At state of delivery delay impending, the flow goes to change in DDR,
delivery delay recognized, and sales effectiveness. After state of sales effectiveness,
10
the flow goes back to order booked. Other flows are also presented in Figure 6.
Comparing with sequencing diagram, activity diagram is more similar with
stock flow diagram because they are built by concept of flow whatever data flow or
activity flow. Therefore we consider that activity diagram of UML would be most
similar with stock flow diagram of SD. It’s corresponding with previous researches.
|
Sales Effectiveness j=
Figure 6 Activity Diagram of Simplified Market Growth Model
Finally, programming code of DY NAMO for system dynamics and JAVA for
UML could also be presented in Table 4. Programming codes in Tables is only to
compare the structure of these two different programming languages, and some
1
detail mathematic functions of program would be ignored. Actually, calculation of
differential equations in system dynamics model needs to be calculated by numerical
methods like Euler’s method or Newton’s method. If these numerical methods are
written by programming language, it will be large programs with many lines of
codes. Therefore, for non-linear problems which need to solve complex partial
differential equations, DYNAMO would be easier to program than nommal
programming languages like JAVA, C++ or Visual BASIC. If system dynamics can
be integrated with information systems and DYNAMO can be embodied in
programming languages, many complex non-linear problems can easily to be solved.
12
Table 4 Simplified Market Growth Model Programmed by DY NAMO
Programming Codes by DY NAMO
Programming Codes by JAVA
Sector: Sales
Salesman.K=Sales.J+(DT)(SalesmanHiring.JK)
Salesman=10
SalesBudget=(OrderBooked)
OrderBooked=(Salesman.k)(RevenuefromSales)
RevenuefromSales=10
ExpectedSales=SalesBudget.K/SalemanSalary
SalesmanSalary=2000
SalesmanHiring.KL=(ExpectedSales.k-Salesman.K)/SaleAdjustTime
SalesAdjustTi
ona raOyv, DAF
e=20
Sector: Orders
OrderBooked.k=(Salesman.k)(SalesEffectiveness.k)
OrderEntered.KL=OrderBooked.K
Backlog.K=Backlog.J+(DT)(OrderEntered.JK=OrderComplete.JK)
Backlog=8000
DeliveryRate.K=TABLE(DR,Backlog.k,0,100000,20000)
TDR=0/10000/16000/18500/19500/20000
OrderComplete=DeliveryRate.K
DeliveryDleaylmpending.K=Backlog.K/DeliveryDelayRecognized.K
ChangeinDDR.KL=(DeliveryDelayImpending.k)/TimeforDDR
TimeforDDR=6
DeliveryDelayRecognized.k=DeleiveyDelayRecognized.J+(DT)(ChangeinDDR.JK)
DeliveryDelayRecognized=2
SalesEffectiveness.K=TABLE(TSE, DeliveryDelayRecognized.K,0,6,1)
TSE=400/390/350/290/210/150/100
4>rpzroaonwmradnarz2zrap
Class Sales {
Public float Salesman=10;
Public float SaleAdjustTime;
Public float ExpectedSales;
Public float SalemanSalary=2000;
Public float RevenuefromSales;
Public SalesmanHiring(ExpectedSales, Salesman, SaleAdjustTime)
{ FUNCTION(ExpectedSales,Salesman,SaleAdjustTime);
Return Salesman; }
Class Orders {
Public float OrdersBacklog;
Public float DeliveryDelayRecognized=2;
Public float Backlog=8000;
Public float TimeforDDR;
Public float DeliveryDelaylmpending;
Public float DeliveryRate;
Public float SalesEffectiveness;
Public OrdersEntered(OrderBooked, Baclog) {
FUNCTION (OrderBooked,Backlog);
Return Backlog;}
Public OrdersComplete(DeliveryRate,Backlog){
FUNCTION(DeliveryRate,Backlog);
Return Backlog;}
Public ChangeinDDR(DeliveryDelaylmpending, TimeforDDR,
DeliveryDelayRecognized){ FUNCTION (DeliveryDelaylmpending, TimeforDDR, DeliveryDelayRecognized);
Return DeliveryDelayRecognized;}
13
4. Integrated Software Systems and Development Process for UML
and SD
4.1 Composition of Re-usable System Dynamics Models
Current large software is usually constructed by components which is a tested,
special-purpose software unit. A software component is useful, adaptable, portable,
and interoperable (Peters, 2000). So many large information systems are combined
by components which make systems be reused and be maintained easily.
Some software tools of system dynamics also provide functions reusable models.
Powersim Studio argued their software products support for object-orientation and
reusable models. Users can define a library of individual models for reuse (Hauke,
2001). However, it is still not software components which can be integrated with
existed information systems. In this paper, the concept of reusable software
components for system dynamics model would be proposed. Component diagram of
UML is shown in the right side of Table 6. Components in component diagrams can
mean an executable file, a data file, a picture or a movie. Components are operated or
communicated through interface. In the left side of Table 6, software components of
system dynamics model is be presented. Many existed models can be encapsulated as
sectors and every sector can be connected by auxiliaries. Then, a large model can be
composed by existed sectors.
Table 6 Component Diagrams of UML and SD
Component Diagram of UML Sectors composition of System Dynamics
r-
== to b}------------ ==
Sector]
t
'
f '
Component} | 2’ Interface? Sector3 :
4}—K Any — —
For example, in the basis of Table 6, an integrated new business dynamics model
can be constructed as Figure 6. There were some existed system models which built
for growth and leaming, intemal process, customers, and finance. Someday, a
company wants to build a Balanced Scorecard (Kaplan, 1996) and they can
14
composed a new system dynamics model by existed models which can be regarded
as sectors and connected them by auxiliaries. By an integrated development process
with SD and UML which will be discussed later, they can develop information
systems and system dynamics models synchronically and develop large SD models
by methodologies of object- orientation.
[onan n enna anne :
' 1
Transformations “e—J Growthand | Transformation! \S—| Internal
\ leaming |-, Transformation? Q—} Process to
' ene e enna i
i 1
; '
i 1
; '
' 1
i '
|
i '
| = |
'
i '
|
[— Transformation3
Figure 6 An Integrated System Dynamics Model
4.2 Integrated Development Process
For software development process, three principles identified in IEEE
1074-1995 standard are:
1. Requirements, to decide what a system must do.
2. Design, to determine how a system computes.
3. Implementation, to produce source code, documentation, and tests
These three principles are fundamental processes for software development.
After that, many software development processes are also proposed which included.
waterfall software process model, incremental life cycle model, spiral life cycle
model, prototyping model, extreme programming, and Rational Unified Process.
Every development process has its advantage and disadvantage and software
engineers decide which one to choose depends on the characteristics of projects.
Rational Unified Process is the development process designed for UML, and it
includes initial planning, business modeling, requirements, analysis & design,
implementation, test, and deployment (Kruchten, 1998). In this paper UML is the
methodology to model business process and therefore Rational Unified Process is
also the development process be chosen.
In System Dynamics modeling process, Sterman (2000) listed them as Problem
Articulation (Boundary selection), Dynamic Hypothesis, Formulation, Testing, and
15
Policy Formulation & Evaluation were the iterative modeling process presented as
Figure 7.
1. Problem Articulation
fo (Boundary Selection) s,
5. Policy Formulation 2. Dynamic
& Evaluation Hypothesis
Ne 3. Formulation
4, a
Figure 7 Iterative System Dynamics Modeling Process (Sterman, 2000)
The phase of problem articulation is to decide what the problem is and horizon
of time. In dynamic hypothesis, develop maps of causal structure based on current
theories. Then, formulate a simulation model. In phase of testing, test robustness
under extreme condition and sensitivity. At last, policy design and evaluation is
scenario specification (Sterman, 2000).
After compared with SD development process and Rational Unified Process, we
found that there were many similar phase between these two processes. An
integrated system development process with system dynamics and UML is presented
as Table 7.
Table 7 Integrated Development Process
Rational Unified Process | System Dynamic Diagrams of SD and
Modeling Process UML
Initial Planning Problem Articulation Use Case
Business Modeling Use Case
Dynamic Hypothesis
Requirements Use Case, Casual
Analysis Fommulation Class, Object,
Collaborative, Causal
Design Component, Sequencing,
Activity, State,
Stock-Flow
16
Implementation Deployment, Stock-Flow
Testing
Test Stock-Flow
Deployment Policy Formulation & Stock-Flow
Evaluation
In this integrated development process, UML and System Dynamics models are
analysis and design synchronously. Through simulation by system dynamics, we can
evaluate final result of information system before it be deployed and correct it
iteratively.
For test and deployment phase of software development process, there is no
suitable diagrams in the process. If we want to test software, we need to complete
the whole program or at least, we need to have a prototype. Nevertheless, stock-flow
diagram can become the role of test and deployment. System dynamics can also do
scenario testing for information systems.
4.3 Integrate Information Systems
Base on the integrated development process, the integrated information system
proposed. by Tu (2003) would be implemented. Traditional business information
systems included Transaction Process System (TPS), Management Information
System (MIS) and Decision Support System (DSS). After we combined UML and
SD, the result simulated by system dynamics can be the data of decision support
system. System dynamics software can also get data from database to overcome
shortages of current system dynamics soft wares.
Transaction Management Decision
Process Information Support
System System System
Figure 9 Integrated Information Systems (Tu, 2003)
By the perspective of system structure, traditional information system would be
17
composed by data-bases, application programs, and user interface. An integrated
structure adds system dynamics into the position of application programs which can
develop like Table 4 and Table 5.
Application
| Programs
User
Interface
Figure 10 Integrated Software Structure
5. Condusion and Suggestion.
System dynamics needs to be integrated with business information systems, and
it can be really applied popularly in the future. UML is the most popular modeling
language to describe business process and we compare it with system dynamics.
After compared with system dynamics and UML, we found that system dynamics
can solve the problems of time delay and system complexity which traditional
information systems can’t. Therefore, concept of integrated information system with
UML and SD has been proposed.
In order to build the integrated information system, system dynamics diagrams
and UML diagrams can be drowned and program codes can also be transformed
synchronically. By the concept of components, a new and large system dynamics
model can be constructed by existed models which regard as sectors. Therefore
system dynamics model can also be developed by methodologies of software
engineering. Design and implementation of physical integrated information will be
next topic in future research.
Reference
Booch, G Rumbaugh, J. and Jacobson, I. 1999. The Unified Modeling Language
User Guide. Addison Wesley Longman, Inc.
Forrester, J. 1975. Market Growth as Influenced by Capital Investment. Collected
papers of Jay Forrester. Portland, OR: Productivity Press, 11-132
Hammer, M. 1990. Reengineering Work: Don’t Automate, Obliterate. Harvard
Business Review 68(4): 104-113
Hauke, U. and Berende, K. 2001. Powersim’s Business Modeling and Simulation
18
Tools Are Built Right In to SAP SEM.
http://www. powersim.com/common/pdf/sap-powersim.pdf
Jacobson, I. 1992. Object-Oriented Software Engineering: A Use Case Driven
Approach. Addison Wesley Publishing Company
Kaplon, RS. and Norton, DP. 1996. The Balanced Scorecard: ‘Translating Strategy
into Action. Harvard Business School Publishing Corporation.
Kruchten, P. 1998. The Rational Unified Process: An Introduction. Addison Wesley
Longman, Inc.,
Marshall, M. 2000. Enterprise Modeling with UML. Addison Wesley Longman, Inc.
Ould, MA. 1995. Business Processes. John Willey & Sons
Object Management Group. 1991. The Common Object Request Broker:
Architecture and Specification.
Peters, JF and Pedrycz, W. 2000. Software Engineering: An Engineering Approach.
John Wiley & Sons, Inc.
Porter, MJ. 1985. Competitive Advantage. Free Press: New York
Reed, Jr PR. 2002. Developing Applications with JAVA and UML. Pearson
Education Inc.
Rumbaugh, J., Jacobson, I. and Booch, G. 1999. The Unified Modeling Language
Reference Manual. Addison Wesley Longman Inc.
Schach, SR. 2002. Object-Oriented and Classical Software Engineering. 5” ed.
McGraw-Hill Companies, Inc.
Senge, PM. 1990. The Fifth Discipline: The Art and Practice of the Learning
Organization. Doubleday Currency.
Sommerville, I. 2004. Software Engineering. 7” ed. Addison Wesley
Sterman, J D. 2000. Business Dynamics. McGraw-Hill Companies Inc.
Tignor, W. 2003. Stock and Flow, and Unified Modeling Language Relationships.
Proceedings of 21"" International Conference of the System Dynamics Society.
Tignor, W. 2004. System Engineering and System Dynamics Models. Proceedings of
22" International Conference of the System Dynamics Society
Tu, YM and Wy, TF. 2003. A Study of Connect Dynamic Data Source to Improve the
Simulate Technique for the System Dynamics. Proceedings of 21" International
Conference of the System Dynamics Society.
19
