Environmental Engineering Reference
In-Depth Information
8.2 State of the Art Research
It has been highlighted that meeting the grand challenge of Autonomic
Systems will involve researchers in a diverse array of fields, including sys-
tems management, distributed computing, networking, operations research,
software development, storage, AI, and control theory, as well as others [ 42 ].
There is no space here to cover all the excellent research underway, so this
section will discuss a selection of the early reports in the literature of state-
of-the-art efforts in AC [ 147 ].
8.2.1 Machine Design
A paper in [ 70 ] discusses affect and machine design [ 101 ]. Essentially, it
supports those psychologists and AI researchers who hold the view that affect
(and emotion) is essential for intelligent behavior [ 139 , 140 ]. It proposes three
levels for the design of systems:
1. Reaction: The lowest level, where no learning occurs, but where there is
an immediate response from the system to state information coming from
sensory systems.
2. Routine: The middle level, where largely routine evaluation and planning
behaviors take place. The system receives inputs from sensors as well as
from the reaction level and reflection level. At this level of assessment,
there are results in three dimensions of affect and emotion values: positive
affect, negative affect, and (energetic) arousal.
3. Reflection: The top level of the system receives no sensory input and has
no motor output; it receives all inputs from below. Reflection is a meta-
process where the mind deliberates about the system itself. Operations at
this level look at the system's representations of its experiences, its current
behavior, its current environment, etc.
Essentially, the reaction level sits within the engineering domain,
monitoring the current state of both the machine and its environment,
and produces rapid reaction to changing circumstances. The reflection level
may reside within an AI domain, utilizing its techniques to consider the
behavior of the system and learn new strategies. The routine level may be a
cooperative mixture of both the reactive and reflection levels.
8.2.2 Prediction and Optimization
A method known as Clockwork provides predictive autonomicity by regulat-
ing behavior in anticipation of need. It involves statistical modeling, tracking,
and forecasting methods [ 127 ] to predict need and is now being expanded
to include real-time model selection techniques to fulfill the self-configuration
Search WWH ::




Custom Search