Completable Planning
Completable planning is an approach combining classical, deliberative
planning techniques with execution-time decision-making within an
explanation-based learning framework. The combination of prior
deliberation and runtime decision-making enables a completable planner
to successfully construct and execute actions despite imperfect domain
knowledge and prior information, while the integration of learning
enables a completable planner to automatically adapt its plans to
particular problem distributions for improved performance.
Selected Publications
M. T. Gervasio
"
An Incremental Curative Learning Approach For Planning With Incorrect
Domain Theories,"
Ph.D. Dissertation, University of Illinois at Urbana-Champaign,
1996. (Also appears as Technical Report UIUCDCS-R-96-1972
(UILU-ENG-96-1733), Department of Computer Science, University of
Illinois at Urbana-Champaign.) [671K, gzip'd]
M. T. Gervasio and G. F. DeJong,
"
An Incremental Learning Approach to Completable Planning,"
Proceedings of the Eleventh International Conference on Machine
Learning, New Brunswick, NJ, July 1994. [276K]
M. T. Gervasio and G. F. DeJong, "
A Completable Approach to Integrating Planning and Scheduling,"
Proceedings of the First International Conference on Artificial
Intelligence Planning Systems, College Park, MD, June 1992. [78K]
M. T. Gervasio, "
Using Qualitative Reasoning in Proving Achievability,"
Proceedings of the 29th Institute of Electrical and Electronics
Engineers Conference on Decision and Control, Honolulu, Hawaii,
December 1990. [79K]
M. T. Gervasio, "
Learning General Completable Reactive Plans,"
Proceedings of the Eighth National Conference on Artificial
Intelligence, Boston, MA, August 1990, pp. 1016-1021. [183K]
M. T. Gervasio and G. F. DeJong, "
Explanation-Based Learning of Reactive Operators,"
Proceedings of the Sixth International Workshop on Machine Learning,
Ithaca, NY, June 1989, pp. 252-254. [77K]
M. T. Gervasio and G. F. DeJong, "
Completable Planning: A Curative Learning Approach to the Imperfect
Theory Problem,"
Working Notes of the 1996 American Association for Artificial
Intelligence Spring Symposium on Planning with Incomplete Information
for Robot Problems, Palo Alto, CA, March 1996, pp. 13-16. [119K]
M. Gervasio and G. DeJong, "
Completable Scheduling: An Integrated Approach to Planning and
Scheduling,"
Working Notes of the 1992 American Association for Artificial
Intelligence Spring Symposium on Practical Approaches to Scheduling
and Planning, Mark Drummond, Mark Fox, Austin Tate and Monte Zweben
(ed.), NASA Ames Research Center, May 1992, pp. 122-126. [158K]
M. Gervasio and G. DeJong, "
Goal-Oriented Sensing in Completable Planning,"
Working Notes of the 1992 American Association for Artificial
Intelligence Spring Symposium on Selective Perception, Palo Alto, CA,
March 1992. [135K]
HOME