Preface We have to look for routes of power our teachers never imagined, or were encouraged to avoid. -
Thomas Pynchon, Gravity's Rainbow
p.3 The process of scientific discovery is, in effect, a continual flight from wonder. - Albert Einstein,
Autobiographical Notes
p.4 The field of complex systems must direct its "flight from wonder" toward discoveries that "make the
wonderful and complex understandable and simple."
p.8 Kenneth Boulding summarized science as consisting of "testable and partially tested fantasies
about the real world." The science of complex systems is not a new way of doing science but rather one in which new
fantasies can be indulged.
p.11 Regardless of the approach, the quest of any model is to ease thinking while still retaining
some ability to illuminate reality.
p.35 We begin with a discussion of the basics of scientific modeling... despite its importance,
most people do not want to discuss it, and no matter how much you read about it, it just doesn't seem the same when you actually
get around to doing it... Effective models require a real world that has enough structure so that some of the details
can be ignored.
p.39 The success of a particular model is tied to its ability to capture the behavior of the real
world.
p.36,37 One of the best models that we encounter in our daily experience is the road map...
Maps are valuable for a variety of reasons. One reason is that they leave out a lot of unnecessary details.
In doing so, they... allow us to focus on the questions that we most care about... Good maps not only allow us to
predict key features of the world, but they also enable us to discover new phenomena.
p.43 Creating a model is much like trying to solve a brain teaser. Finding such solutions
is often an extremely difficult task... Yet, once discovered, the answer has strong intuitive appeal and appears all too obvious.
p.44 Much of the focus of complex systems is on how systems of interacting agents can lead to emergent
phenomena... The usual notion put forth underlying emergence is that individual,
localized behavior aggregates into global behavior that is, in some sense, disconnected from its origins.
p.50 When interactions are not independent, feedback can enter the system. Feedback fundamentally
alters the dynamics of a system. In a system with negative feedback, changes get quickly absorbed and the system
gains stability. With positive feedback, changes get amplified leading to instability.
p.59 Currently, the predominant tool in economic theorizing is the development of mathematical models derived
from a set of first principles... we can use these models to make predictions... These tools provide both a ready set of simplifications
for understanding the world and a process by which the implications of these simplifications can be derived in a consistent
way.
p.62 All tools are designed to simplify some task. If the task we face corresponds with
this simplification, then the tool will be of value.
p.62 By attacking problems on numerous fronts, breaches in nature's walls inevitably appear.
Though it may be difficult to predict on which front the walls will first crumble, openings on one front are likely to lead
to progress on another.
p.65 the goal of theory is to make the world understandable by finding the right set of simplifications.
Modeling proceeds by deciding what simplifications to impose on the underlying entities and then, based on those abstractions,
uncovering their implications. The types of computational models we wish to focus on here are those in which the abstractions
maintain a close association with the real-world agents of interest, and where uncovering the implications of these abstractions
requires a sequential set of computations involving these abstractions.
The property of close association is often known as "agent-based" modeling... A marginally better
term might be modeling using "agent-based objects" versus "abstraction-based objects"
p.69 To create models that go beyond our initial understanding, we need to incorporate frameworks
for emergence. That is, we need to have the underlying elements of the model flexible enough so that new, unanticipated
features naturally arise within the model.
p.72 The ability of a model to generalize is linked closely to its inherent frameworks for emergence.
p.72 Computational models are often thought to be brittle, in the sense that slight changes in one area
can dramatically alter their results... For good modeling we need to keep in mind the brittleness of our tools and actively
work to avoid producing theories that are too closely tied to any particular assumption.
p.83 Many of our existing analytic tools avoid an emphasis on dynamic processes and focus on equilibrium
states... In natural systems, however, equilibria are usually associated with the death of the system. The conditions that
favor equilibrium analysis are likely the exception rather than the rule in many complex adaptive social systems. If so, the
techniques that we traditionally use to analyze these systems may be like trying to "understand running water by catching
it in a bucket."
p.84 Social scientists have often recognized the importance of dynamic analysis but have been very constrained
by their tools. According to Von Neuman and Morgenstern (1944, 44) [JLJ - Theory of Games and Economic Behavior ],
in their seminal work on game theory, "We repeat most emphatically that our theory is thoroughly static. A dynamic
theory would unquestionably be more complete and therefore preferable. But there is ample evidence from other branches
of science that it is futile to try and build one as long as the static side is not thoroughly understood."
p.84 In situations where equilibria are nonexistent or transient paths are long, understanding the
dynamics is critical.
p.94,95 Given the limits of information processing, agents must actively ignore most of the potential information
they encounter... We suspect that a full analysis of how agents selectively attend to information will provide some interesting
scientific opportunities... agents may develop ways, such as statistics, to summarize the flow of incoming information so
that it is easily stored and used... agents may... influence the actions of others... the reality is that most agents exist
within a topology of connections to other agents, and such connections may have an important influence on behavior.
p.96 In social systems, we may want the agent behavior to aggregate in such a way that the system achieves
some goal.
p.101 Right Concentration [JLJ - a model construction principle ] is the focus of the model - namely, it
requires the model to be just sufficient to capture the phenomenon of interest. Models always have contexts, and what works
well in one context may fail in another... Models of complex systems phenomena should be simple, not complicated.
This point often seems to get confused and twisted in various ways, but the point of modeling - whatever the target - is to
simplify an otherwise overly complex world. Thus, even when the resulting behavior is complex, the underlying model
should be simple.
p.114 Modeling any system is often an exploratory process that requires both induction and deduction. You
begin by making a simple set of assumptions and see where they lead. From this experience you attempt to create better models
or deduce more exact results.
p.139-140 the structures necessary for delicate behavior require an underlying system that is rich
in possibilities. In essence, we need a quivering system that will fall into the right state with only a
gentle tap. In such a system, an improper tap can lead to very unpredictable results.
Adaptive systems have to deal with the tension between the benefits of achieving precise behavior
and the cost of increased system fragility. One hypothesis is that adaptive systems will have a bias toward emphasizing
simple structures that resist chaos over more complicated ones that handle difficult situations.
p.141 Far better an approximate answer to the right question, which is often vague, than the exact answer
to the wrong question, which can always be made precise - John Tukey, Annals of Mathematical Statistics
p.177 adaptation can alter the critical behavior of the system... With adaptive agents, the system configures
itself in a way that mitigates the overall risk by preventing criticality from emerging. In essence, the adaptive
actions of the individual agents lead the system away from the critical regime and more toward what an omniscient designer
attempting to balance risk and stability would create.
p.214-215 A key contribution of complex systems has been a better appreciation of the power and
mechanism of emergence... Perhaps many features of social systems are the result of self-organization... Models of
emergence also provide insights into the robustness of the underlying system, as the essence of emergence requires
entities to be able to maintain their core functionality despite what are often radical changes from both within and without.
Using emergence ideas, we can begin to understand the robustness of systems such as markets, cultures, and organizations like
firms and political parties.
p.220 Chess has fascinated players in its modern version for over half a millennium...
chess... has a well-defined game tree that, in theory, we could work our way through and develop an optimal strategy. If our
cognitive abilities were just a bit higher, all the fuss about chess might be a bit embarrassing... Like humans, computers
are unable to generate the entire game tree for chess except toward the end of the game. Therefore, programs must rely on
various heuristics... and other various means to decide on their moves.
p.236 Inherent in many complex adaptive social systems is a degree of robustness... which details of the
agents matter in terms of maintaining the system's coherence? Second, we can consider the robustness of the entity to
perturbations in the environment. What does it take for a system to persist in the face of external changes?
Alternatively, we could frame the question as uncovering the factors that make a given system brittle [JLJ - the opposite
of robust ].
p.246 Making sure that your model has just enough of the right elements and no more is the most
fundamental practice for any kind of modeling, and computational work is no exception... Scientific modelers must
aim for... simplicity and clarity. Modeling is like stone carving: the art is in removing what you do not need.
It is often easy to recognize a simple, well-formulated model after the fact, as such models
have strong intuitive appeal... Good models strip phenomena down to their essentials, yet retain sufficient
complication to produce the needed insights.
p.249 Often it is possible to create computational models with simple, flexible frameworks that "get filled
in" by the computation... it is the evolution of the system that fills in the details. A well-designed framework puts very
few a priori constraints on the model, and thus the outcome is rich in possibilities. Such frameworks provide enough flexibility
so that the model can explore areas that were not fully anticipated by the researcher... Note that even the simplest of frameworks
can result in outcomes that are difficult to understand.
p.254 Scientific judgments in this area [computational modeling] should focus not on the computer per
se, but on the quality and simplicity of the model, the cleverness of the experimental design, and the new insights gained
by the effort.
