xiv Our world is dynamic, but the human mind is ill-equipped
for comprehending, assessing, and predicting dynamic processes... It is possible to construct computer simulation
models of dynamic processes and systems, and it is possible to use these models for gaining a better understanding
of a system and its dynamics, and of the options available for coping with the ensuing problems... [This] book is meant for
students of all disciplines - dynamic systems are found everywhere. [JLJ - perhaps even in game positions]
p.1 the computer can track the multitude of implications of complex
relationships and their dynamic consequences much more reliably than the human mind.
p.2-3 Our task will be to present the procedures and methods for
modeling and simulating the dynamic systems that exist all around us. We will be dealing with a common approach
suitable for all disciplines... This is possible because the properties and behavior of dynamic systems are
determined not by their physical structure and appearance but by their system structure and processes...
In addition to helping in the solution of particular problems, systems analysis, modeling, and simulation therefore
also help us in gaining a much more general understanding of dynamic systems and their behavior
p.4 A model is sought that would be able to simulate behavior and
to provide hints concerning necessary actions to avoid inadmissible or even dangerous developments.
The product we are interested in is therefore a reliable substitute
description that can help one understand a real system and its expected behavior.
p.4 A second approach [JLJ - at the simulation of behavior] is to
attempt an explanation of behavior by modeling the actual processes of the real system. In this case much has to be known
about the system itself: Of what parts is it composed? How are they connected? How do they influence each other... the emphasis
is on the description of structure and processes.
p.5-6 Modeling and computer simulation of a dynamic system may lead
to a new understanding that would not follow directly from the original knowledge about the system... Computer simulation
is also used to better manage existing dynamic systems.
p.23 On closer look, all systems are dynamic systems
p.45 In order to show that the model system can represent the original system
well enough for the model purpose, validity must be demonstrated with respect to four different aspects: behavioral validity,
structural validity, empirical validity, and application validity.
p.45 Model construction always means simplification, aggregation,
omission, and abstraction.
p.87 The model captures the major processes... in considerable detail. It
links all these processes together in a complex dynamic system. The dynamics that the model produces are the result
of the dynamic interactions of the many individual processes. If the individual processes have been correctly described,
and if the complex relationships are faithfully represented in the model, we should expect reasonable agreement with
developments in the real world.
p.91 Recognition of the behaviorally relevant system structure and
its representation in the influence diagram constitute the first phase of model development... these results are
neither sufficient nor reliable for understanding exactly how a system may behave over time.
Modeling of dynamic systems therefore also requires a close look
at the individual elements and their particular role and function in the system. Introducing their specifications
into the influence diagram, we obtain the simulation diagram which now contains all the information necessary to program and
compute the simulation... We will find that the notion of "state," and the description of dynamic systems by their
"state equations," are useful concepts for developing the simulation model and for analyzing the system which it represents.
The development of the state equations is therefore at the heart of the modeling process.
p.103 Clever choice of the state variables can lead to particularly simple
system descriptions.
p.155 Programming new simulation software for each individual application
therefore does not make sense.
p.230 The set of criteria that are relevant for the evaluation of system
development will be referred to as orientors. This term is used to make clear that we are dealing with criteria that systems
(or their managers) use to orient their decisions and actions regarding the system. Orientors are aspects, terms,
or dimensions (like "freedom") that designate important criteria or qualities of system survival and development but not the
degree to which they must be satisfied.
To evaluate system development with respect to the relevant orientors,
indicators describing the system state must be mapped on the orientors in order to compare the actual system
state with the corresponding goal values in "orientation space."
In the evaluation of system development, three different types of
normative criteria play a role:
1. Constraints that limit state variables, system variables,
final system states, the duration of processes, or control inputs to permissible ranges.
2. Quality measures that allow differentiating "better"
from "worse" state developments and may also allow the search for "optimal" solutions.
3. Weights that allow an aggregated evaluation of several
quality measures of different degrees of importance must be applied simultaneously.
p.232 Where do the criteria for the evaluation of system development
originate? They obviously have an influence on the kind and extent of system observation or model construction... The selection
criteria is determined on the one hand by the immediate interests of the investigator, and on the other hand
by his/her understanding of the system and the problem.
p.236 If a system is to survive and thrive in an environment characterized
by these fundamental properties: [normal environmental state, scarce resources, variety, variability, change, other
systems] then it must either have been forced to pay attention to these properties of the environment during its evolution,
or it must be designed to adequately cope with them.
p.238-242 [orientors listed] existence... effectiveness... freedom
of action... security... adaptivity... regard for other systems [Muller lists this as coexistence]
p.247 If a system is to be viable in the long run, a minimum satisfaction
of each of the basic orientors must be assured.
p.247 In the assessment and orientation of system behavior we therefore
deal with a two-phase assessment process where each phase is different from the other:
1. First, a certain minimum satisfaction must be guaranteed separately
for each of the basic orientors. As long as this is not the case, the system's attention will have to be focused
on the remaining orientor deficits.
2. Only if the required minimum satisfaction of all basic orientors
is guaranteed is it permissible to use a weighted sum of surplus orientor satisfactions to obtain an overall quality
or satisfaction index which can then be maximized.
Here again we come across constraints (which must be observed strictly and
individually and cannot be balanced one against the other); quality measures (where surplus satisfactions can be computed
in terms of a common "currency" - for example, "utility" - and be balanced one against the other); and weights (determining
the relative importance of different contributions in aggregated quality measures).
p.274 As natural systems evolve, their environment forces them to
respect the basic orientors.
p.286 It became obvious that the fundamental structure of a dynamic
system and its embedding in its system environment pose certain requirements that must be fulfilled if the system is to remain
viable, functioning, and adaptive. The "basic orientors" that can be derived from these requirements
can serve as a framework for assessments when evaluating alternative development paths, searching for better policies
or even "optimal" solutions, trying to stabilize unstable systems, and designing new systems.
