This presentation will provide an overview of several state of the
art constraint-based scheduling and optimization techniques developed at
the Computational Intelligence Research Laboratory (CIRL) at the University
of Oregon. These techniques are being applied to real-world applications
including shipbuilding, CD manufacturing and aircraft assembly where they
have been able to significantly reduce the scheduling time and increase
the size of problems that can be handled. The presentation will focus primarily
on the ``SqueakyWheel Optimization'' and ``Schedule Pack'' algorithms and
will provide comparisons of these techniques with existing scheduling approaches.
In addition to the algorithm descriptions, several small demonstrations
will also be given. The presentation will conclude by drawing future
research directions and describing potential application areas.
Vita
Brian Drabble received
his undergraduate degree for the University of Staffordshire, UK in 1984
and his Ph.D. in artifical intelligence from the University of Aston, UK
in 1988. On completion of his Ph.D. he joined the Planning and Scheduling
Group of the Artificial Intelligence Applications Institute (AIAI)
at the University of Edinburgh. During his time at AIAI he lead a number
of projects investigated the development of frameworks for practical
planners and schedulers such as the O-Plan System, which was funded under
the DARPA/Rome
Laboratory Planning Initiative. In addition to these projects,
he worked with several industrial clients including Toshiba, Hitachi, BAe
and the European Space Agency.
Brian Drabble joined the Computational
Intelligence Research Laboratory (CIRL) at the University of Oregon
in January 1997 with the main responsibility of transfering CIRL's basic
research in planning and scheduling to clients in industry, commerce and
the military. He is a Senior Research Associate and was Director of CIRL
from July 1999 through July 2001. Brian Drabble is a principal in On
Time Systems, a commercial spin-out of CIRL that delivers scheduling
and optimization systems to various organisations including Lockheed Martin,
General Dynamics, Bath Iron Works, Air Combat Command and Air Mobility
Command.
Brian Drabble's research interests include high-speed scheduling systems,
resource optimization, temporal reasoning and real-time systems.