Linking System Dynamics Model to Data Modeling
Yi-Ming, Tu, Wang Wei-Y oung
Department of Management Information System
National Sun Y at-sen University, Kaohsiung, Taiwan, R.O.C.
wyoung@ mis.nsysu.edu.tw
Abstract
Traditional top-down design of databases is unable to fulfill managerial information
needs. A database is supposed to contain all the information required for decision
making. However, the information requirements are not only uneasy to identify but
also variable in the dynamic complex environment. The value of a database to support
management may decline. The research is concerned about how to define
management information requirements clearly and effectively. In the paper, system
dynamics' models are combined into the information requirement analysis stage of
database planning. Both operational and managerial information requirements can be
extracted and derived by transforming system dynamics models to data models. In
addition to support database design, system dynamics’ model also provide managers
the opportunity to aware how they used to perceive and interpret decision tasks.
1. Introduction
Database has become a core technology in organizational computing. A
database contains most of organization’s digitized data. With those digitized data,
daily transactions and routine works can be done more efficiently, and the
effectiveness and timeliness of management and control activities can be improved
also. To satisfy an organization’s information needs, system analysts use top-down
planning and bottom-up designing approaches. Top-down planning is to extract an
enterprise data model that will be the blueprint for later development and integration
of each separate database. In the enterprise model, future organizational
development is supported. Besides, bottom-up designing approach is used to
decompose the enterprise data model and to implement physical databases
according to their priorities..
General top-down analysis procedure(Mcfadden et al., 1993) starts from
organization's strategic planning. Then, managers transform the result of strategic
planning into sub-goals and more detailed activities. According to the nature and
contents of those activities, data items to be processed are identified. In addition,
relationships among data items are also derived from operational characteristics and
constraints. The final result in this information requirements analysis is called data
model.
The resulting data model is expected to provide information for operational
activities and to support managerial decisions. However, it does not work well as its
expectancy. Organization’s strategies and goals may change frequently; database
can not satisfy mangers' information needs; information systems are inflexible,
etc.(King, 1978; Lederer & Sethi, 1988; Lederer, 1991). In essence, because
managers’ decision tasks are unpredictable, managers cannot identify their
information requirements well. Furthermore, because most of management
environments are dynamic and complex, cognition structure inconsistency happens
to managers frequently. In such circumstances, managers feel that they need more
information to resolve their cognition structure inconsistency, but they may not
know exactly what information they need. The original information needs may be
distorted or lost in the top-down transformation process.
The paper attempts to use system dynamics to support database planning.
Information requirements are defined more effectively, and the priorities to develop
sub-data models are also determined.
2. Data Model
The content of a data model can be divided into two parts. One is data items, the
other is the way data items are organized. The data model is one kind of data
structure representations. Its purpose is to provide data to support each level of
organizational activities, such as operational and managerial activities(Dehayes,
1990). In practice, different levels of organizational activities should have different
information requirements and data models. Most of the database designs are built on
the operational-level data model, while data model or information requirements of
other organizational levels are ignored. Several kinds of data models are proposed
and developed, for example, hierarchical data model, network data model, relational
data model, and object-oriented model, etc. Because relational database is most
popular nowadays, the paper is supposed to use it to illustrate the research.
A relational data model is composed of entity types and relationship types.
Entity types are objects that physically exist and they can be described by their
attributes. Relationship types are relationships among entity types. Entity types are
obtained from information requirements analysis, while relationship types are
primarily extracted from operational rules and constraints.
When a database is unable to satisfy users’ information needs, there might be
two problems with the data model. One is that data model does not contain
appropriate entity types for required data items. This occurs when the information
requirements are uncertain and difficult to identify. The other problem is that the
relationship types defined in the data model are incomplete or incorrect. Both of the
two problems happen more frequently in the dynamic and complex environment.
3. The Role of System Dynamics Modeling in Database Planning
Database planning has two phases. The first phase is to derive information
requirements from an organization's strategic planning and operational activities.
The second phase is to transform the information requirements developed earlier
into a data model and to design the physical database. The quality of a database is
determined mostly by the success of first phase.
Morecroft(1984) indicates that if a model is to be effective in support of
strategic planning, it must be (1) down from the pedestal of the infallible black box
to occupy a more modest position as a complement to the thinking and deducing
powers of management; (2) a vehicle for extending argument and debate; (3)a
generator of opinions, not answers, and managers must be encouraged to challenge
and debate model conclusions. System dynamics’ models are able to satisfy those
requirements. Prior research has suggested several ways to implement system
dynamics models in strategic planning(Morecroft, 1984, 1985; Morecroft et al.,
1991; Risch et al., 1995). The paper extends prior research to support managerial
information requirements analysis.
In addition to managerial requirements, operational information requirements
can also be derived from system dynamics’ models. In terms of isomorphism, the
characteristics of operational processes are reflected in system dynamics’ models.
Hence, actual operational processes can be elicited from the system dynamics’
models. According to those processes, operational information requirements can be
acquired
With system dynamics’ models, managerial information requirements can be
more explicit and operational information requirements can be obtained more easily.
And multiple-level transformation processes are eliminated. Techniques to
transform system dynamics' models to data models will be discussed in detail.
4. The Linkage between Data Model and System Dynamics Model
As described above, data model can be established from the transformation of
system dynamics’ models and the analysis of operational processes. This section is
devoted to illustrate how the transformation processes proceed.
In system dynamics models, level is the major information provider. If we look
at levels more closely, tangible and intangible levels are present. For examples,
people, money, equipment, and orders (Forrester, 1961)are tangible levels, while
stress and cognition are intangible levels. In the data model, only tangible levels are
included as data items(entity type). Further, through different kinds of computation
processes, extended tangible level can be obtained from original tangible levels. If
the data model support management function, then these original tangible levels
should be included into the data model. Rate equations point out what users’ view
are. Users’ view is a subset of data model, which is information needed when
making a decision. Because users’ view is extended information from levels, it is
used just for user interface design.
However, a data model should include more than data items obtained from the
transformation of system dynamics’ models. To support those policy designed in
system dynamics’ models, a data model should provide information for operational
processes. Operational information requirements can be obtained from the analysis
of operational processes. A fter obtaining managerial information requirements from
the system dynamics model, analysts have to find out basic operational processes
indicating by the model. With traditional requirement analysis methodology, those
processes are analyzed and the conceptual data model design is finished.
Besides, in terms of technical consideration, it may be inefficient to query some
users’ views involving many complex data transformations. In order to increase the
database’s performance, such users’ views can be materialized into the database.
5. Conclusion
The effectiveness of a database is dependent on the completeness of its data
model. However, in variable environments, traditional database planning is unable
to acquire stable and concrete information requirements. Furthermore, information
losses and biases arise more frequently. The research attempts to use system
dynamics models as a vehicle to induce and extract organization’s information
requirements. Three advantages emerge by doing so. First, it is more effective in
dealing with strategic issues. Second, it is easier to define more clear and more
stable information requirements for operational and managerial support. Finally, it
can help information users to emphasize on the assumptions and interpretations they
made. With system dynamics’ models, realization of what's valuable information
and what information managers really need is deepened.
Reference
Dehayes, D. W., J. A. Hoffer, E. W. Martine, and W. C. Perkins, MIS for Managers,
New Y ork, N.Y., MacMillian, 1990.
Forrester, J. W., Industrial Dynamics, MIT Press, 1961, pp. 82.
King, W. R., “Strategic Planning for Management Information Systems,” MIS
Querterly(2:1), Mar. 1978.
Lederer, A. L. & V. Sethi, “Application and Implementation Critical Dimensions of
Strategic Information System Planning,” Decision Science, Vol. 22, 1991.
Lederer, A. L. & V. Sethi, “The Implementation of Strategic Information Systems
Planning Methodologies,” MIS Querterly, Sep., 1988.
Mcfadden, F. R., Jeffre A. H., Database Management, Third edition,
Benjamin/Cummings Publishing Company, Inc., California, 1991, pp. 6.
Morecroft John D. W., D. C. Lane, and P. S. Viita, “Modeling Growth Strategy in a
Biotechnology Startup Firm,” System Dynamics Review, Vol. 7, No. 2, Summer
1991.
Morecroft, John D. W., “Strategy Support Models,” Strategic Management Journal,
Vol. 5, 1984.
Morecroft, John D. W., “The Feedback View of Business Policy and Strategy,”
System Dynamics Review, Vol. 1, No. 1, Summer 1985.
Risch, J. D., L. Troyano-Bermudez, and J. D. Sterman, “Designing Corporate Strategy
with System Dynamics: a Case Study in the Pulp and Paper industry,” System
Dynamics Review, Vol. 11, No. 4, Winter 1995.
CC BY-NC-SA 4.0