Conference article

PSI Team: Planning Method and Main Principles of the Agents Architecture

Alexander Kozhushkin
Program Systems Institute, Russia

Download article

Published in: RobocCup-99 Team Descriptions. Simulation League

Linköping Electronic Conference Proceedings 4:24, p. 110-115

Linköping Electronic Articles in Computer and Information Science vol. 4 4:24, p. 110-115

Show more +

Published: 1999-12-15

ISBN:

ISSN: 1650-3686 (print), 1650-3740 (online)

Abstract

The paper describes the PSI team developed at Program Systems Institute of Russian Academy of Science. We propose a dynamical refinement method of plans with defaults for the construction of software agents playing soccer. The architecture of our agents is composed of two main parts: the set of elementary plans and the planning system. The planning system uses on-line plan refinement procedure for constructing plans with elementary ones. There are some default conditions of normal continuation associated implicitly with each plan. The purpose of the refining process is to add new actions to the initial plan in order to maintain or recover normal continuation of the plan.

Keywords

No keywords available

References

[1] T. Dean; S. Kambhampati. Planning and Scheduling. CRC Handbook of Computer Science and Engineering (1996).

[2] M. Ginsberg. Modality and Interrupts. Journal of Automated Reasoning (14)1 (1995); pages 43-91.

[3] M. Ginsberg 1995. Approximate Planning Arti cial Intelligence Journal (76)1-2(1995); pages 89-123.

[4] M. Ginsberg; H. Holbrook. What defaults can do that hierarchies can’t. Fundamental Informaticae (21) (1994); pages 149-159.

Citations in Crossref