|
Architecting
the Cognitive Enterprise: Towards Learning and Self-Adaptive Organizational
Networks
Cite as:
Mostashari, A.” Architecting the Cognitive
Enterprise: Towards Learning and Self-Adaptive Organizational Networks”,
COMPASS Whitepaper 2009-01, Stevens Institute of Technology, February 2009
Dr. Ali Mostashari
The
ability of an enterprise to adjust the allocation of its resources and the
type or magnitude of its outputs to rapidly changing external and internal
needs and requirements is critical. It hinges on the capacity of the
enterprise to sense and/or forecast the pertinent changes, contextualize
the information, make timely decisions that affect resource allocation and
product configuration, monitor the impact of these decisions and learn for
improved future decision-making.
There
have been many studies on the concept of Agile Enterprises, which can
anticipate and respond to changes in the market in a rapid manner [1-3].
However the Agile approach puts the emphasis of response to change
to human centered deliberations on the initial design of enterprise
structures and processes to allow for agility in response to change. Also
most of the literature is limited to heuristics and qualitative frameworks
that cannot be robustly applied, tested and validated for different
enterprises under a variety of conditions.
The goal
of this research is to explore the concept of “Cognitive Enterprises” as a
new paradigm for architecting learning and self-adaptive enterprises that
can dynamically adjust their resources, configuration and outputs to meet
the constantly changing external and internal needs and requirements. The
term Cognitive refers to the ability of the enterprise to “learn to
adapt its behavior based on experience”. Hence, Cognitive Enterprise relies
on autonomous, intelligent observation, trade-off analysis and action
accompanied by learning to allow an Enterprise to increasingly become
better at reconfiguring itself in response to change.
The Cognitive Enterprise
 The concept of Cognitive Enterprise is
based on the “Cognitive Radio” paradigm, developed by one of the co-PIs of
this research (Dr. Joseph Mitola III) in 1999 [4], and expanded and refined
in subsequent years by him and other researchers [5-11]. Cognitive Radio
was defined as “radios that can sense their environment, including spectrum
use and the presence of other such radios and then adapt their
transmission/reception to avoid interference could implement a dynamic
control method for solving the interference problem.”[6]
The
potential of a “Cognitive” approach for Enterprises is significant. It
heralds the evolution of an enterprise that can sense its environment,
including other competing and collaborating enterprises and adjust its
resource allocations, internal configuration and outputs in a way that
optimizes its value delivery. The study of Cognitive Enterprises also
builds on insights from the fields of cognitive linguistics, game theory
and enterprise architecture to explore an interdisciplinary approach to
architecting enterprises that are capable of learning and self adaptation.
As such it represents a new analytic approach to fundamentally defining
enterprises based on their ability to dynamically respond to change.
Based on the insights from the Cognitive
Radio Literature [5], we define a Cognitive Enterprise as having the
following capabilities:
1.
Sensing individual internal and external changes
2.
Perceiving the overall picture that these changes
represent
3.
Associating the new situation with past experienced
situations and acting accordingly if similar
4.
Planning various alternatives in response to the
change within a given response timeline
5.
Choosing course of action that seems best suited
to the situation
6.
Taking Action by adjusting resources and outcomes to
meet new needs and requirements
7.
Monitoring and Learning from the impact of capabilities 1-6
From the
definition it follows that every enterprise could exhibit these
capabilities in different degrees. Each of these capabilities is used in an
Enterprise process that directly corresponds to it. The chain of the seven
resulting processes constitutes the full cognitive process cycle for the
Enterprise for any given set of changes. The focus of this Cognitive
Enterprise research is on cyber-enabled close-to-real time usage of these
capabilities in order to meet rapid and consecutive changes in needs and
requirements.
In
structuring this research, we build the groundwork for future research to
address the following fundamental issues with regards to Cognitive
Enterprises:
· What different levels of
cognition can exist within enterprises and how can they be measured/
evaluated?
·
How do enterprises sense and perceive changes in their
environment? Who or what within the organization does the sensing and perceiving?
·
How is information from the sensing domain translated
into/understood as contextual information that supports resource and output
decisions within the enterprise?
·
To what extent is the decision-making performed by a cognitive
data network and when does the social network (the network of the
enterprise’s human resources) enter the decision-making dynamics?
·
How do the different sensing agents within the enterprise
negotiate resource usage?
·
What types of enterprise architectures best support Cognitive
Enterprises?
·
How do Cognitive Enterprises learn from previous
experiences?
Even a partial answer to
these questions will constitute a significant contribution to our
understanding of how organizations respond to change.
|
