Our
vision is to develop design and optimization methods that change
the way design is done in practice. In the new design environment,
the full power of optimization can be made accessible to all practitioners,
and brings about immense new possibilities. We believe that the coalescence
of salient developments over the past decade makes this vision essential
from both industry and academic perspectives.
PHYSICAL
PROGRAMMING: Physical programming plays
a key role in several aspects of the research performed at the laboratory.
In short, physical programming is an effective and efficient optimization
method for the masses and the expert alike. As an example of its early
use, a few years ago five undergraduate students with no prior knowledge
of optimization entered a national ASME design competition, and designed
a nonlinear control system for an industry manufacturing system. They
used physical programming to identify optimal designs that could have
been obtained only through the most advanced theoretical control analyses.
The student team emerged as competition winners. The physical programming
method has since been used in a comprehensive list of practical applications.
DESIGN:
We view the design process as a decision-making activity that uses
the foundational principles of engineering, economic, and management
disciplines to create a system or design that will address some perceived
societal need.
OPTIMIZATION:
We view optimization as a powerful tool that can critically impact
the design of all systems whose performance can be quantified. This
encompasses practically all engineering and economic systems.
MULTIDISCIPLINARY:
We use the term multidisciplinary from a human perspective (collaborating
with people from different disciplines), from a disciplinary
perspective (simultaneously addressing distinct disciplinesin our
analyses and optimization), and from a computational perspective (computational
and modeling infrastructures are heterogeneous).
NEW
POSSIBILITIES: The computer and information technology revolution
of the past decade has made it possible to use optimization to design
engineering and non-engineering systems in radically different ways.
It is now possible to use reasonable computational and human resources
to explore the design space of systems (i) large and small, (ii) of
different temporal and spatial scales, (iii) involving disparate disciplines
interacting in harmoniously competitive ways, and (iv) involving engineering,
economic, and other non-engineering considerations.
CHALLENGES:
The salient challenges include: (i) The development of effective and
efficient optimization methods that are accessible to the average
engineer, and non-engineer for non-engineering applications. Physical
Programming addresses these issues. (ii) The development of practical
methods for the inclusion of uncertainties to yield robust systems
and designs. (iii) The development of modeling methods of the proper
computational resolution for both real-time interaction, and high-fidelity
performance evaluation. (iv) The development of real-time visualization
methods for design and optimization.
In
addition, we maintain leading research in the areas of control, dynamics,
and structural dynamics.
RESEARCH
ACTIVITIES: Research activities include:
(i)
The development of optimization methods for concept selection,
where the consequences of decisions have the most impact (Christopher
Mattson),
(ii)
The development of the Normal Constraint Method for effectively exploring
the design space (Amir Ismail-Yahaya),
(iii)
The examination of critical flaws of commonly used industry
methods for concept selection, and the development of new methods
that effectively circumvent these deficiencies (Anoop Mullur),
(iv)
The development of economically competitive production planning
methods for industry application (Aniela Maria),
(v)
The development of optimization methods for chromatographic processes
(Deepak Nagrath, Prof. S.M. Cramer, and Prof. B.W. Bequette),
(vi)
The development of optimization methods for leaner structural approaches
for housing (Amy Farina and Prof. S. Van Dessel),
(vi)
The development of optimal product family design methods (Michael
Martinez, and Prof. T. Simpson),
(vii)
The development of robust design methods (J. G. Sundararaj
and Prof. W. Chen),
(viii)
Tradeoff analysis and decision making in design (R. V. Tappeta and
Prof. J. E. Renaud),
(ix)
Subsystem selection (Prof. K. Lewis), and
(x)
Development of a new underwater propulsion system invented
by Prof. Messac that is poised to challenge conventional systems,
both in energy efficiency and other critical performance metrics (Professors
Hirsa, Koratkar, and De).
Other
research projects include (i) The development of Control-Structure
Integrated Design methods, (ii) The development of dynamics and optimization
methods for the Shuttle Orbiter Stabilized Payload Deployment System
- used for massive and voluminous payloads, (iii) Flexible body dynamics
for the Space Station Mobile Transporter.
COLLABORATIONS: We collaborate with industry and academic leaders
nationally and internationally, and we actively interact with colleagues
here at RPI as well, notably in engineering, chemistry, and architecture.
STUDENTS:
All graduate students typically engage in archival journal publication
and presentation of their work at national conferences. Students learn
from each other through interaction and collaboration as much as they
learn from the faculty members.
