Problem domains
While broad applicability is one measure of an idea's usefulness,
measurable and significant performance improvements in specific areas
of interest always provide the most compelling demonstrations of a
technique's worth. CIRL has traditionally evaluated our
representation and reasoning techniques against real-world application
domains, both as a way of providing objective evaluations of various
techniques and because real-world problems often reveal interesting
features that can be exploited to make progress in understanding how
to solve the underlying computational problem. This approach has
often produced the world's best solvers in particular application
areas:
- The bridge-playing program GIB has set an entirely new standard
for computer performance in this game.
- WARP, our Worldwide Aeronautical Route Planner, is capable of
finding optimal flight paths using actual weather and performance
data for various aircraft.
- Our scheduling work on problems provided by large manufacturing
firms has produced schedules of substantially better quality than
existing techniques. This includes both the techniques currently in
use on the factory floor and techniques being developed by other
academic groups.
Parent areas:
Applications
Subareas:
Bridge
Route planning
Scheduling
Back to the CIRL Overview Page.