Systems, Man & Cybernetics - October 2015 - 19

We now can pursue
research questions
that past luminaries-
the shoulders on
which we stand-
could not have
imagined as tractable.

best to decompose phenomena of interest, represent these phenomena, and
recompose these representations into
overall models of complex systems and
enterprises. At the most basic level is
the concern about consistency of assumptions. Are inputs,
outputs, process, and parameters based on shared definitions of variables, units of measure, and coordinate systems? These must be handled correctly for anything to
make sense.
Much more difficult is the problem of entangled
states. When simultaneously solving multiple sets of
equations, is the solution an equilibrium solution or is
it premised on indefensible assumptions, which could
lead to unstable responses? For example, traffic flow
is typically modeled using partial differential equations. In contrast, driver decision making is usually
modeled with agent-based models. Thus, we quickly
encounter situations where traffic flow affects driver
decisions that, in turn, affect traffic flow. Our assumptions about this causality has an enormous impact on,
for example, congestion pricing rates that are intended to discourage dr ivers from ventur ing onto the
roads at peak demand times. Drivers may interpret
high prices as signals that demand is high and consequently seek the higher-priced rather than the lowerpriced solution.
We conducted surveys in four industries regarding how
they address these model composition issues. The industries included automotive, building equipment, commercial
aerospace, and semiconductors. Everyone agreed about
the problems of model composition. They were particularly concerned with avoiding model-induced design errors.
Their concern was that the models used to answer the last
set of questions would provide misleading or outright
wrong answers to the next set of questions. We need to
understand these modeling issues at a much deeper level.
The question is not simply one of how to get disparate
models to compute together. Instead, we need a more disciplined level of understanding of how different representations do and do not fit together, as well as the
consequences of forcing the fit so that computation is feasible but meaning is lost.

Wiener Symposium in San
Diego, California. In my talk,
I expanded on Norbert Wiener's classic book Cybernetics:
On Control and Comm-
unication in the Animal and the Machine. I broadened
the perspective to "Control and Communication in Animals, Machines, Organizations, and Societies," and I
empathized that this broader perspective leverages
many ideas that were articulated long ago by Wiener,
Shannon, Weaver, and many others. The difference is
that we now have enormous computational capabilities,
large-scale interactive visualizations, and massive data
sets including numeric data, text, and images. We now
can pursue research questions that past luminaries-
the shoulders on which we stand-could not have imagined as tractable. We need to leverage this legacy to
make a difference.
About the Author
William B. Rouse (wrouse@stevens.edu) is the Alexander Crombie Humphreys Chair in the School of Systems
and Enterprises at Stevens Institute of Technology and the
director of the Center for Complex Systems and Enterprises. He is also a Professor Emeritus and a former chair of
the School of Industrial and Systems Engineering at the
Georgia Institute of Technology.
References

[1] E. Pate-Cornell, W. B. Rouse, and C. M. Vest, Perspectives on Complex Global
Challenges: Education, Energy, Healthcare, Security, and Resilience.
Hoboken, NJ: Wiley, 2016.
[2] W. B. Rouse, Modeling and Visualization of Complex Systems and
Enterprises: Explorations of Physical, Human, Economic, and Social
Phenomena. Hoboken, NJ: Wiley, 2015.
[3] W. B. Rouse, "Human interaction with policy flight simulators," Appl. Ergonom.,
vol. 45, no. 1, pp. 72-77, Jan. 2014.
[4] W. B. Rouse and K. R. Boff, Organizational Simulation: From Modeling
and Simulation to Games and Entertainment. Hoboken, NJ: Wiley, 2005.
[5] W. B. Rouse and N. Serban, Understanding and Managing the Complexity
of Healthcare. Cambridge, MA: MIT Press, 2014.
[6] M. J. Pennock and W. B. Rouse, "The challenges of modeling enterprise systems,"
Proc. Int. Symp. Eng. Syst. (CESUN), Hoboken, NJ, 2014.
[7] M. J. Pennock, W. B. Rouse, "Why connecting theories together may not work: How

Prospects
In 2014, I was a panelist at the IEEE International
Conference on Systems, Man, and Cybernetics Norbert

to address complex paradigm-spanning questions," in Proc. IEEE Int. Conf. Syst.,
Man, and Cybern., San Diego, CA, 2014.

O c tob e r 2015

IEEE SyStEmS, man, & CybErnEtICS magazInE

Table of Contents for the Digital Edition of Systems, Man & Cybernetics - October 2015