Download Transition from Simple, Complicated to Complex Systems

Editorial
Transition from Simple, Complicated to Complex
Systems
Mustafa M. Aral 1
Abstract
As our understanding of micro components of engineering and science fields improve, our macro
definition of these systems are also advancing. New concepts and definitions introduced in redefining the
systems we work with has catapulted the engineering and science fields to a more sophisticated level,
where these fields can no longer be characterized as simple or complicated systems, but at macro scales
these systems are characterized as complex systems. When we compare this development with the
progress made in the methods of analysis and the solution techniques that are used in these fields we see
that there is a significant scientific lag between the two. This lag leaves most of the redefined complex
problems currently unsolvable. If our goal is to provide solutions to complex systems, it is clear that new
concepts and new methods need to be introduced for the analysis and solution of these problems as well.
These advancements have also altered the decision making process after one provides a successful
solution of the complex system. In this Editorial one such concept is outlined for the purposes of initiating
a discussion on these new perspectives that are necessary in the in the analysis of complex problems.
Introduction
As scientists specializing in the broad field of engineering and its scientific applications, whether we
are an engineer, scientist, ecologist, health scientist or a member of another profession that is linked to
any of these fields, we know that the systems we work with exhibit a behavior that cannot be
characterized as being stable and/or near steady state. Most of these systems are in a constant flux that
are producing significant changes in the overall behavior of the system over time. Unfortunately, even
with this awareness, we still have a tendency to treat these systems as systems which are relatively simple
or complicated, and ignore the broader behavior of the system that is constantly undergoing and
responding to change. What is the cause of this complacency, what are the symptoms and outcomes?
Probably the tendency to consider problems in the realm of simple or complicated applications and ignore
complexity is stemming from the overwhelming uncertainty associated with the broader description of
1
Professor and Director, Multimedia Environmental Simulations Laboratory, School of Civil and Environmental
Engineering, Georgia Institute of Technology, Atlanta, Georgia, USA.
1
these problems as complex systems and characterization of the interlaced linkages between the
components of these systems. However, if progress is our goal, this should not deter us from seeking
alternative methodologies to address the problem in a more meaningful manner, as well as seek methods
of solution to these complex systems using new analysis techniques which will eventually lead to more
meaningful management decisions.
In this Editorial our focus is on three principal concepts we use in engineering. The first concept is
associated with the definition of the problem at hand, which can be identified as a simple, complicated
or a complex problem. The second concept is associated with the purpose of the analysis, which can be
defined as a control, mitigation or survival purpose. The third concept is associated with the decision
making process, which can be characterized as crisis management, risk management or adaptation. As
we move from simple to complicated and finally to complex problem definitions our purpose of analysis
moves from control to mitigation and to survival, and the decision making process moves from crisis to
risk management and finally to adaptation. In this sense, complexity theory and analysis presents us with
a sophisticated tool for long term planning and management, but brings about more questions than
answers in the analysis techniques. As our understanding of complex and chaotic behaviors improve, the
new stability definitions that are introduced become an essential element of the study of long term
consequences of the decisions we are making today.
As engineers and scientists we are trained, and probably are very good at, in understanding and
solving simple or complicated problems in our narrow fields of specialization. This approach requires us
to further specialize in the area of research we are engaged in, and focus our attention to that specific
field alone. Unfortunately, most problems in engineering and science fields cannot be studied in isolation
and within the narrow fields of our specialization. Further, we are also well aware of the fact that the
problems we work with cannot be characterized as simple or complicated systems within the narrow fields
of our specialization, but they need to be characterized as complex systems. Components of complex
systems cannot be studied in isolation from each other since complex systems are dynamical systems with
many components which are originating from a wide range of contexts. This feature of complex systems
will require the participation of a team of specialists from several fields in formulating the problem and
also the generalists to show us the links between the specializations involved.
A complex system can be defined as a dynamical system that is comprised of many components
having relatively many interlaced relations among them, so that the behavior of each component is
influenced by the behavior of the other components, Figure 1. The traditional approach in most
engineering or scientific analysis on the other hand, is so far based on the premise that the systems we
are working with are simple or complicated but not complex and the interlaced linkages can be ignored
or feedback mechanisms are assumed to have negligible effect on the overall behavior of the system. One
may characterize this as the typical unidirectional behavior of a complicated system, Figure 1. In this
Editorial we will not devote much space to the discussion of simple systems since they are very elementary
applications.
The procedures used to solve complicated problems, which also effects the decision making process
after the problem is solved, is based on the premise that the management decisions can be designed to
2
control the effects of unidirectional change that is imposed on the system. This approach may yield
satisfactory outcomes for systems that are assumed to be stable and/or near steady state. In these
applications, the premise we have is that the change observed on the system is possible to control. We
may characterize this approach as the traditional approach of analysis of complicated problems. If the
complicated system has many variables we may also move to the concept of mitigation as an alternative.
Unfortunately this is not a good strategy to analyze complex systems, where the concept of control does
not have a meaning. Similarly the decision making process moves from crisis management to risk
management and finally to adaptation processes as we move from complicated to complex systems.
b. Complex system with feedback
a. Complicated unidirectional system
Figure 1. Complicated vs. Complex Systems
Environmental Engineering Perspective
For the purposes of this Editorial, given the space limitations we have, we may focus on environmental
problems and the methods of analysis of these problems. Traditionally, in these problems our focus is on
the evaluation of the effect of one environmental stressor on one entity via a single or multiple
environmental pathways. This does not consider the role of a combination of several environmental and
more importantly non-environmental stressors on many entities and the effect of change produced on
each entity based on the change observed on the other. Here a “stressor” can be defined to be a chemical,
physical, biological stressor and the “entity” can be defined to be a societal or environmental entity that
maybe impacted by the stressor. Overall, some of these stressors may contribute to adverse effects on
entities, and others might increase resilience of the system under study such as the social support
structures if one of the entities involved is the society. How these factors interact with each other and
influence the outcome and the behavior of the overall system has significant implications on both the
interpretation of the behavior of the system, the analysis methods used in the solution of the problem
and also the decision making process after the solution is achieved. For example, a pollution level at a
source and exposure to that pollution as the pollutant migrates in various environmental pathways may
be totally harmless if the pollution levels are low, if the pathways does not interact with communities or
cause permanent harm to the environment. However, given the same definitions and assumptions for this
problem, the health effects outcome of exposure to this low level stressor maybe amplified due to societal
stresses that it creates on the community or a segment of the community such as different age groups. In
3
order to protect populations from adverse health effects, an understanding of the complex interactions
between various stressors among various segments of populations and physiologies is critical and this
analysis should also include the appropriate treatment of the uncertainty associated with the data
whether it is of statistical, stochastic or heuristic origin.
In environmental engineering, when complicated systems are analyzed using traditional methods of
analysis, and after partial success or even maybe complete success of the narrow goals of the study is
achieved, we always find ourselves trying to justify how successful the project or research outcome was
when we try to articulate the outcome and try to extrapolate its implementation within the broader
framework of the complex system. That is when engineers fail to be convincing and that is when mistakes
are made in the environmental sciences field. The main reason for this failure is that the original study
always tends to ignore the complexity and the effect of the change the system is experiencing due to
changes occurring within its components. Thus, in that interpretation, the projection simply becomes a
speculative extrapolation of the narrow knowledge gained in the narrow field of the application. The
reason behind this deficiency is because the integrated analysis of complex systems was not the focus
area of the traditional definition of the problem and methods used in its solution at the first place.
Whereas, we are well aware of the fact that the surroundings in which we live in, function and enjoy life
as human beings can only be described in terms of the principles embedded in the complex system theory.
There are other examples of complex environmental systems that is currently on the agenda of
researchers, such as the problems associated with climate change based human-environmental systems
(Aral, 2010; NSF, 2009; USEPA, 2010). For example, the premise in complicated climate analysis is that,
climate change may create warmer temperatures which may increase air and water pollution and this in
turn may have adverse human health effects. Thus, if we control temperature increases on earth the
problem can be solved. This is a typical unidirectional approach for the solution of a complicated problem
and a crisis management decision that is based on the concept of control. We are well aware of the fact
that climate processes can only be characterized as a complex system and they do not respond to change
in a smooth and predictable manner. A stressed or perturbed complex human-climate-environmental
system may suddenly shift from a seemingly stable steady state to a state that is difficult for it to reverse
its path and turn to its original state after the change takes its toll on the system. Further, not only these
complex systems in nature and society do change, but over long periods they also change how they
change. This is a very important characteristic of complex systems. Thus, change is possible to control
premise of our traditional models and analysis strategies becomes impractical in the analysis of complex
systems. In complex systems the individual components of the system and the interactions between them
may lead to large scale changes and behaviors which are not easily predicted from the knowledge and the
analysis of the behavior of the individual components. This concept is in contrast to the perspective of a
world that is in near equilibrium and/or in steady state as they are defined in most of our current humanclimate-environment applications. Parallel to this concept the decisions made should not be based on
control instead we need to adopt the decision process of adaptation, which is embracing and adopting to
change. These are large scale problems and require significantly different concepts and thinking out of
the mainstream line of thought for solutions to be meaningful and also useful.
4
In the human-climate-environmental system example, the outcome of an irrecoverable change, which
is going to be the outcome of the recommended strategies based on the complicated problem analysis,
will be loss of life or loss of the social fabric of our complex societies. That is a tipping point we would like
to avoid as much as possible at all times in contrast to other fields in which some failures may be tolerated.
Thus, in the human-climate-environment field we find ourselves in a more precarious position in decision
making. In the case of complicated problems the decision making process can be characterized as crisis
management or for more advanced cases risk management. In the case of complex systems the decision
process is characterized as adaptation. These two decision making processes require significantly different
frameworks as well as methods of analysis. Examples of this nature are many, where certain thresholds
are nearby and they are critical. Decisions made based on risk analysis may yield a more soft landing type
of an impact as opposed to the abrupt impacts of a crisis management decision. However, adaptation is a
more realistic management strategy for dynamic complex systems. From this discussion it is becoming
clear that not only these problems cannot be solved through traditional analysis of complicated problems
but they may be addressed adequately by an extended risk based analysis and adaptation to change as
we treat the system as a complex system.
Resilience Thinking
In analyzing complex systems, in addition to understanding how dynamic systems behave, it is
also important to understand the return to equilibrium concept. For such analysis, the methodology that
embraces the concept of resilience and its components; latitude, resistance, precariousness and panarchy
are very important (Gunderson et al., 2002; Holland et al.; 1995, Kauffman, 1993; Neubert et al., 1997;
Gros, 2008). Resilience of a complex system is defined as the capacity of the system to absorb disturbances
(stresses) while undergoing change as it essentially retains its function, structure and identity, and
response state. In essence, in the context of human-climate-environment system, a resilient human
behavior would be the one that would have the capacity to respond to stresses introduced to the state of
the system such as temperature increase, sea level rise, social stressors and other potential stress factors
without exhibiting failure modes such as adverse health outcome or deterioration of the social fabric of
the society. The loss of resilience may lead to more vulnerable states in which even minor further
disturbances may cause a significant shift in the state of the system and moves the system into another
state that is difficult or even impossible for it to recover from. Thus, vulnerability is the flip side of
resilience concept and occurs when a system loses its resilience and becomes vulnerable to change that
previously it could have absorbed. The resilience concept has four components which are quantifiable
using basic engineering and science based tools — Latitude (L), Resistance (R), Precariousness (Pr), and
Panarchy (Pa). Latitude is defined as the maximum amount the system may change before it loses its
ability to recover; Resistance is the ease or difficulty of enacting a change on the system; Precariousness
is the current trajectory of the complex system, and how close it currently is to a threshold which, if
breached, makes recovery difficult or impossible or moves the system to another state; Panarchy is an
indicator to measure how the above three attributes are influenced by the states and dynamics of the
other systems that comprise the overall complex system at scales above and below the scale of interest.
In this manner, when all stressors are included in the resilience landscape (Figure 2), the overall system
analyzed will be an integrated complex system.
5
~
Re(Y)
x~2
L
Pr
R
x~1
Figure 2: Resilience landscape of a complex system and potential stressed
system trajectories for a system of two state variables.
Since human perception or response is an important component in our concept of resilience in
human-climate-environment methodologies, we should also include the quantifiable concept of
adaptability and transformability to our tools of analysis in a heuristic sense. These measures are related
to the capacity of making desirable system basins of attraction in the resilience landscape wider and/or
deeper, while shrinking undesirable states and the introduction of new stability landscapes by introducing
new components and ways of making the overall system work more harmoniously. In essence, in this line
of thinking, the stability dynamics of the linked systems of human, climate and environment merge from
the three complementary attributes of resilience, adaptability, and transformability (Boccara, 2004). In as
much as the resilience of humankind to climate can be defined using basic engineering and scientific
principles, we also acknowledge the fact that resilience based human-climate-environmental systems
should be also linked to social resilience concept. Here we introduce the definition of Social-Ecological
System (SES) which is an integrated system of ecosystems and human society with reciprocal feedback
and interdependence. This concept emphasizes the humans-in-nature perspective. This is a concept
associated with social resilience which is a measure of the ability of a community to cope with the stress
imposed on them by the presence of a stressor in their community. Notably, social resilience differs
fundamentally from natural or engineered system resilience since it exhibits the capacity of humans to
anticipate and plan for the future and this plays an important role in human-climate-environmental
studies. Unfortunately, although these concepts are iterated in the literature many times over in the
ecology field, their computational counterparts to develop the resilience landscape and the return to
equilibrium analysis are at their infancy as far as the environmental scientists would like to define them
in our computational models and applications. Thus, there is a lot of work to be done in this line of thought
and research, but mostly these methodologies can be linked to minimum energy solutions through
optimization analysis of our engineering methodologies. This concept of linkage of complicated systems
and their solution and their adaptation to complex systems and their solution is demonstrated in the
solution for a complex savannah management system with the involvement of a few entities (Guan, Zhang
and Aral, 2013) and health-climate-environment system (Aral, 2010).
6
Direction of Complex Systems Research
The U. S. National Science foundation Advisory Committee for Environmental Research and Education
has released a report that advocates a significant change in direction in the way environmental system
research and education will be perceived by this agency in the future (NSF, 2009). In this report, the NSF
committee is advocating that physical and life scientists, engineers, educators and social scientists must
work collaboratively to understand and evaluate the behavior of complex systems under the stresses
imposed on the system. An important theme of this report is that scientists need to consider that the
human-climate-environmental systems that involve human component may be approaching thresholds
of irreversible change given the nature of the complexity of the stressors. This is also due to the
unpredictable nature of human transformability and adaptability characteristic which also renders the
analysis to be rather complex. The concepts embedded in the resilience analysis, as briefly described
above, may shed some light to the problems posed in the NSF report which needs to be pursued by the
scientific community although the computational aspects of this approach are overwhelming (Aral, 2013;
Aral, 2015).
Conclusions
The quantifiable resilience thinking may yield an actionable set of observations and practices that is
based on the broad understanding of complex human-climate-environmental systems. This approach
does not assume or require that the system studied has to be in equilibrium or near equilibrium at all
times, nor it is controllable. For the previous paradigms of environmental management, precise
understanding of the system was needed and the policy decisions made relied on the accuracy of this
understanding and related predictions. This is a concept that has been shown to fail in several applications
in the past. Currently, the mathematics of resilience thinking is at its infancy. The deterministic analysis
mode that is currently used in resilience analysis is also not sufficient to analyze the resilience of complex
human-climate-environmental systems. However, the idea is promising and many applications in complex
systems analysis that involve humans and policy are shifting to models that include the resilience
concepts. This approach offers a broader understanding of possible system behavior and the effects of
stochastic and fuzzy human intervention on this behavior. We do not need to resist change, but instead
we need to embrace it by changing our perspective of analysis from the study of complicated problems
to the study of complex problems following the line of thinking introduced in resilience analysis.
References
Aral, M.M. 2010. “Resilience Analysis of Climate Change Effects on Water Quality and Health,” Editor, A.
Baba, Proceedings of the NATO Advanced Research Workshop on Climate Change and Its Effects on
Water Resources, Izmir, Turkey.
Aral, M.M. 2013. Climate Change and Human Population Dynamics. Journal of Water Quality, Exposure
and Health: Special Issue on Climate Change, Vol. 5, No. 4, pp. DOI: 10.1007/s12403-013-0091-5.
7
Aral, M.M. 2015. Climate Change and Persistent High Temperatures: Does it Matter? Frontiers in
Environmental Science, Section on Interdisciplinary Climate Studies. DOI: 10.3389/fenvs.2014.00045.
Boccara, N. 2004. Modeling Complex Systems. Springer-Verlag, New York, ISBN: 0-387-40462-7.
Gros, C. 2008. Complex and Adaptive Dynamical Systems. Springer Publications, London New Yok, ISBN:
978-3-540-71873-4.
Guan, J., Zhang, A. and Aral, M.M. 2013. An Optimization Approach for Sustainable Development Planning
of
Savanna
Systems.
Journal
of
Arid
Environments,
Vol.
98,
pp.
60-69:
DOI:10.1016/j.jaridenv.2013.07.013.
Gunderson, L.H. and Pritchard, L. 2002. Resilience and the Behavior of Large-Scale Systems. Island Press,
London.
Holland, J.H. 1995. Hidden Order: How Adaptation Builds Complexity. Addison-Westley, Reading, MA.
Kauffman, S. 1993. The Origins of Order. Oxford University Press, New York.
Neubert, N.G. and H. Caswell, 1997. Alternatives to resilience for measuring the responses of ecological
systems to perturbations. Ecology 78(3): 653.
NSF. 2009. Transitions and Tipping Points in Complex Environmental Systems. A Report by the NSF
Advisory Committee for Environmental Research and Education, Washington D.C. 60p.
USEPA. 2010. Climate Change - Health and Environmental Effects, Web site last accessed on August 20,
2016, http://www.epa.gov/climatechange/effects/health.html
8