|
Università degli studi di Udine
|
Artificial Intelligence
Laboratory
|
|
Università degli studi di Udine
|
Goals
Results
Principal Investigators
This research project has two major goals.
The first goal is to show the
appropriateness of the multimodeling approach for representing physical
systems, which has been already been succesfully employed for diagnosis,
for engineering design tasks, in particular conceptual design.The
Multimodeling approach supports the representation of several types of
knowledge about an artifact such as structural, behavioral, functional and
teleological knowledge. Moreover, it provides a general and theoretically
founded framework for organizing and using multiple representations in a
cooperative way for problem solving. Because design specification is
usually expressed in a language far remote from solution description, this
feature is used to support the conceptual translation as from goals, into
functions, then into the artifact's behavior, and finally, into its
structure the design effects.
The second goal is to develop a general
theory for task adaptive multistrategy design that aims at dynamically
combining a range of different design strategies such as transformational
design, design from first principles, and prototype based (or case based)
design, in more powerful application systems that can take advantages of
the strenghts of each approach and thus can be potentially applied to a
much wider range of practical design problems.
Current research work focuses on techniques for run-time assessment and
prediction of solution quality during conceptual design.
Previous page
AI Lab