what-when-how
In Depth Tutorials and Information
interactions. Its overall emergent behavior is difficult to predict, even when subsys-
tem behavior is readily predictable.” (Sussman, 2003) Sussman also defines three
types of complexity in systems: behavioral (also called emergence), internal (also
called structural), and evaluative (Sussman, 2003).
Behavioralcomplexity
arises when the emergent behavior of a system is difficult
to predict and may be difficult to understand even after the fact. For instance,
the easiest solution to traffic congestion seems to be to build new highways. New
highways, however, cause additional traffic by attracting “latent transportation
demand” due to the increased attractiveness of private autos, thus leading to more
congestion in the long run.
Internalorstructuralcomplexity
is a measure of the interconnectedness in the
structure of a complex system, where small changes made to a part of the system
can result in major changes in the system output and even result in systemwide
failure. A good example of this type of complexity is the side effect of chemother-
apy, which, in addition to destroying cancerous cells, also suppresses the immune
system of the body, resulting in death by infection in cancer patients.
Evaluative complexity
is caused by the existence of stakeholders in a complex
system and is an indication of the different normative beliefs that influence views
on the system. hus, even in the absence of the two former types of complexity, and
even if one were able to model the outputs and the performance of the system, it
would still be difficult to reach an agreement on what “good” system performance
signiies. his type of complexity is one of the primary motivators for engaging
stakeholders in systems modeling and policy design and is an essential aspect of
such systems. here are many diferent criteria to value particular outcomes in a
sociotechnical system. Which criteria are used to evaluate outcomes, and how they
are measured, have to be determined by the consensus or overwhelming majority
agreement of the stakeholders. Otherwise, the valuation can be considered that of
the experts and decision makers alone. Some of the social and economic valuation
approaches for outcomes include (Mostashari, 2009)
Utilitarian: his criterion is one of neoclassic economics. Essentially, the goal
here is to maximize the sum of individual cardinal utilities. (W(x) = U1(x) +
U2(x) + ... + Un(x)). Of course, this can only function if U1 is cardinal (and
if the U's are interpersonally comparable).
Pareto optimality: he goal here is to reach an equilibrium that cannot be
replaced by another one that would increase the welfare of some people with-
out harming others.
Pareto eiciency: his occurs when one person is made better of and no one is
made worse off.
Compensation principle: A better-off person can compensate the worse-off per-
son to the extent that both of them are better of.
Social welfare function: Here the state evaluates the outcome based on overall
social welfare, taking into account distributional issues.
