Providing Simulation Models on the Internet
Andreas Gréfler
Industrieseminar der Universitat Mannheim
D- 68131 Mannheim, Germany
Phone: (+49 621) 292-31 40 @ Fax: (+49 621) 292-52 59
e-mail: agroe@is.bwl.uni-mannheim.de
Shared Mental Models in International Firms
That managing organizations and mastering their environment has become a complex (sometimes
chaotic) task is generally accepted.’ Also well known is the insufficiency of the human mind in
dealing with such complexity (which has a variety of reasons, for instance the limited capacity of our
short time memory, Miller 1956, or misperceptions of feedback structures, Sterman 1989). One
prominent way to let people improve their skills to act successfully under these circumstances is to be
busy with computer models (Dérner 1992, pp. 307-309). Through the simulation of organizations and
playing management games individual decision makers and groups are able to gain insights in the
dynamic and chaotic aspects of reality. These so called Microworlds are supposed to be efficient
means to improve the mastering of complex business systems (Senge 1990, pp. 313-316).
In order to accelerate Organizational Learning these tools have to be used in management teams to
»change their shared mental models of their company, their markets, and their competitors“ (deGeus
1988, p. 70). This happens in two steps: firstly, by constructing the underlying formal model which
describes the common view of reality of the involved individuals. This formal model is used to run
actual simulations. Secondly, teams build a shared mental model by playing with the formal model. In
this case, the team members feed-in their hypothetic decisions, run the simulation, and have a look
what is happening in the ,,virtual reality“ (Milling 1995, p. 106).
For international organizations both steps of this procedure can be a problem: although System
Dynamics provides an international ,,language“ for building models and simulations’, technical means
are needed to realize the chances for people from different parts of the world to truly work with one
model. I. e. a ,,place“ is needed where people can bring in their mental models to build a common
formal model and where they can test their mental models by comparing their assumptions with
results from a simulation run.
Through the widespread use of computer networks the possibility to create ,,virtual places“ to allow
people building shared mental models seems obvious. In doing so, new communication technology
not only makes it possible to access static data all over the world, but also enables collaborative
working with dynamic and qualitative information, which mainly constitutes the policy of an
organization (Forrester 1994, p. 58). This technology could be used for internal training as well as to
provide models with the organizations view of the world to external people. Working with the same
models all over the world supports similar mental models of decision makers in international
enterprises. This allows a coherent transformation of strategy to action all over the world.
'This statement is so widely in use that you can find it as the starting point for nearly every book about
management both popular and scientific.
*In a way, that causal diagrams and other graphical explanations of systems can be understood without advanced
knowledge of - for example - the English language. Of course, for the description of the diagram a ,,natural*
language has to be used.
\qo
Basics of Working with Models on the Internet
The rest of the paper only deals with the second step of building shared mental models in an
international organization, i. e. the possibility to provide simulation models in a global scope. In an
earlier step collaborative building of a formal model - which is the base for any simulation experiment
- can be supported by video conferencing systems, electronic mail, electronic whiteboards, etc. In a
broader sense, communication technology can support every of Senge's ,,disciplines“ (Senge 1990), in
particular - but not only - if people have to work together who are geographically separated.
The aim is that people can test their assumptions about effects of their decisions against the results of
a simulation run with the same decisions. The simulation model constitutes the present corporate view
of reality. If they discover substantial differences they either should adapt their mental models (they
learn) or should suggest to change the simulation model (the organization learns). Both are a kind of
Double Loop Learning (Argyris and Schén 1978, pp. 20-26).
The Internet provides a technical infrastructure and in particular the World Wide Web (WWW) gives
opportunities to work on a simulation model in a global scope (Lux and Choi 1995, for a general
description of System Dynamics activities on the Internet). The WWW is built in client/server fashion:
the client (WWW browser at user's computer) displays data which it gets as HTML pages from a
remote (WWW) server over the Internet. The corresponding transportation protocol is called HTTP’.
This mechanism not only allows to access static data but also makes it possible to access models.
Through Common Gateway Interfaces (CGI) input from users can be passed to the core model
(implemented with common simulation languages or tools) somewhere in the world (Ford 1995, for a
description of the WWW and the CGI mechanism). The simulation runs and returns the results (via
WWW server) to the user, who works in the same way as — for instance — with a local ,, Management
Flight Simulator“ (Bakken, Gould and Kim 1995, p. 245, for a short definition of Management Flight
Simulators). The basic logical structure of simulating a remote model therefore looks like this:
Client (user) side Server side
WWW-browser: - HTTP
user input Pn WWW server:
| receiving user input; }~ CGI
transmitting to:model AY Simulation:
simulate with!
| WWW server: mam user input
transmitting results CGI
WWW browser: an to user
displaying results |
Figure 1: Structure of interactive simulations on the WWW
Furthermore, users can also play together: people from different nations can constitute a team and try
to manage a certain game situation. If different teams are able to play, the result is an international
»Corporate Planning Game“.
'HTML means Hypertext Markup Lanugage; HTTP stands for Hypertext Transfer Protocol..
\a)
An Example Simulation Model Provided on the Internet
As an example we can consider a very simple simulation model: the growth of a given population of
mice dependent on their fertility rate and their average lifetime.' There is no difference— in terms of
the needed action to provide it on the WWW — between this simple model and a more sophisticated
one, like a Management Flight Simulator. How can this model be accessed and be influenced via
www?
When somebody wants to run a simulation he or she needs to know the address of a WWW page
which offers this service. The user runs a Web browser with this address and gets a screen, which
offers possibilities to enter needed data. The following screenshot shows this data-entry screen for the
used example:
Andreas Groessler, 11/0496
Screenshot 1: Input form for simulation data
The screen allows to type in the values for those variables which can be influenced. After pushing the
»Simulate!“-button this data is sent back to the server that provides the simulation model somewhere
in the world. There, the actual simulation takes place. Some moments later the user gets the results
back on his or her screen as another WWW page. The results could be in graphical or numerical form,
like it is shown for the example in Screenshot 2.
Providers of models must follow a three step schedule to accomplish this task. They have to:
develop the simulation model itself;
create an HTML page explaining the model and possibilities for user input (see Screenshot 1);
write a CGI application which passes the transfered user input from the WWW server to the
simulation program (which uses the developed model) and generates another HTML page with the
results of the simulation (see Screenshot 2).
'The example model is similar to one which is shipped with Vensim® simulation software, which I use to do the
simulation.
142
‘Screenshot 2: Results' page after simulation run
References
Argyris, Chris and Donald A. Schén: Organizational Learning: A Theory of Action Perspective,
Reading, 1978.
Bakken, Bent, Janet Gould and Daniel Kim: Experimentation in Learning Organizations: A
Management Flight Simulator Approach, in: Morecroft, John D. W. and John D. Sterman (ed.):
Modelling for Learning Organizations, Portland, 1994.
de Geus, Arie P.: Planning as Learning, in: Harvard Business Review, March-April 1988.
Dérner, Dietrich: Die Logik des MiBlingens, Reinbeck, 1992.
Ford, Andrew: Spinning the Web. How to Provide Information on the Internet, London, 1995.
Forrester, Jay W.: Policies, Decisions, and Informations Sources for Modeling, in: Morecroft, John D.
W. and John D. Sterman (ed.): Modeling for Learning Organizations, Portland, 1994.
Lux, Nan and Eric Choi: System Dynamics and the Internet, in: System Dynamics '95, Volume II,
638-645.
Miller, George A.: The Magical Number Seven Plus or Minus Two: Some Limits of our Capacity for
Processing Information, in: Psychological Review, Vol. 63 (1956), pp. 81-97.
Milling, Peter: Organisationales Lernen und seine Unterstiitzung durch Managementsimulatoren, in:
ZfB Erganzungsheft 3/95 Lernende Unternehmen.
Senge, Peter M.: The Fifth Discipline, New York, 1990.
Sterman, John D.: Modeling Managerial Behavior: Misperceptions of Feedback in a Dynamic
Decision Making Experiment, in: Management Science, Vol. 35, No. 3 (March 1989).
\43
