vii This ability to model dynamic systems is already having a powerful influence on teaching and studying
complexity.
p.3 Model building is central to our understanding of real-world phenomena. We all create
mental models of the world around us, dissecting our observations into cause and effect.
p.3 As Heinz Pagels has noted, the computer modeling process is to the mind what the telescope and
the microscope are to the eye.
p.3 Frequently, the phenomena occurring in the real world are multifaceted, interrelated and difficult to
understand. In order to deal with these phenomena, we abstract from details and attempt to concentrate on the larger picture
- a particular set of features of the real world or the structure that underlies the processes that lead to the observed outcomes.
Models are such abstractions of reality. Models force us to face the results of the structural and dynamical assumptions we
have made in our abstractions.
p.4 Key elements of processes and observations can be identified to form an abstract version of real events.
Particularly, we may want to identify variables that describe these events, and outline the relationship among variables,
thereby identifying the structure of the model. Based on the performance and the results of operating or "running" the model,
we can draw conclusions and provide predictions about events yet to be experienced or observed.
p.4 Modeling is a never-ending process - we build, revise, compare, and change models. With each cycle,
our understanding of the reality improves.
p.5 Models help us understand the dynamics of real-world processes by mimicking with the
computer the actual but simplified forces that are assumed to result in a system's behavior.
p.5 It is an elementary preprinciple in modeling that one should keep the simulation simple,
even simpler than one knows the cause and effect relationship to be, and only grudgingly complexify the model when
it does not produce the real effects.
p.5 Computer models are causal in the sense that they are built by using general rules
that describe how each element in a system will respond to the changes of other elements.
p.6 Some of the elements that make up the system for which a model is being developed are referred to as
state variables... Typical nonconserved state variables are price and temperature. System elements that represent
the action or change in a state variable are called flows or control variables.
p.6 Typically, components of the system that is being modeled interact with each other. Such interactions
of system components are present in the form of feedback processes. Feedback processes are said to occur if changes
in a system component initiate changes in other components that, in turn, affect the component that originally stimulated
the change.
p.8 Systems modelers pay special attention to nonlinearities and time lags in their models.
Throughout their lifetime they try to sharpen their perception of nonlinearities and other systems features, and they improve
their skills in modeling them.
p.8 Our mental models are often inadequate to provide a comprehensive perspective on the
many interrelated aspects of systems and to anticipate their behavior.
p.9 Today there are powerful methods and tools of computer modeling available that enable virtually
anyone to develop dynamic models of complex systems
p.9 The very nature of this book and the books of the Dynamic Modeling Series is to help you in
learning how to translate your mental models into rigorously based computer ones and how to engage yourself and others
in a continuous learning process.
p.10 When the models are run, they reveal "normal" system behavior if no interference into
the system takes place, and they may reveal emergent properties of the system.
p.11 while the art of Dynamic Modeling requires that one is skillfully identifying system components and
their interactions, the technique to Dynamic Modeling requires a master plan for the development of the model's structure
- not the details, but the layout of model components... The more interdisciplinary the modeling approach, the more likely
it is that knowledge from different disciplines is brought to bear on the development of the master plan according to which
a model is designed.
p.12 Formation of analogies is one way of dealing with complexity. A great many new insights
are generated by learning something from the structure of behavior of one system, which is well understood, about another
system, of which we have less knowledge. The formation of analogies forces us to choose different systems perspectives. We
identify the structure of one problem and compare it with the structure of another problem. We note their similarities and
their differences.
p.18-19 In general, keep your models simple, especially at first. Compare your results
with measured values where at all possible. Only complicate your model when it does not produce results that predict
the available experimental data within a sufficient level of accuracy.
p.21-22 Principles of Modeling... we think we have learned something general about the
modeling process after many years of trying... We expect you to come back to this list once in a while as you proceed in your
modeling efforts...
- Define the problem and the goals of the model. Frame the questions you want to answer with
the model... Think now: Is my model to be descriptive or predictive?
- Designate the state variables... Keep it simple. Purposefully
avoid complexity in the beginning...
- Select the control variables, the flow controls into and out of the state variables...
- Select the parameters for the control variables... Keep it simple at
the start. Try to capture only the essential feaures...
- Examine the resulting model for possible violations of... laws...
- To see how the model is going to work, choose some time horizon over which you intend to
examine the dynamic behavior of the model...
- Run the model. See if the graph of these variables passes a "sanity test." ...
- Vary the parameters to their reasonable extremes and see if the results in the graph still make sense...
- Compare the results with experimental data...
- Revise the parameters, perhaps even the model, to reflect greater complexity and to meet
exceptions to the experimental results
p.22 a good model enables prediction of the future course of a dynamic system.
p.22 The models developed in this book are all built with the graphical programming language STELLA.
p.43 We can best demonstrate the process of dynamic modeling by describing first the process that
we wish to model and the parts of the model that determine its dynamic behavior. Then we may want to abstract away from a
large number of details that characterize the real system, thereby concentrating on the main driving forces.
p.79 An inability to come with the model at least close to the historical system behavior is a good
sign that you missed some important driver behind the observed dynamics. How closely you wish to replicate past system
behavior depends on the purpose of the model and your motivation.
p.102 At the beginning of the book we discussed the importance of positive and negative feedback processes
for the dynamic behavior of systems.
