Sensory Graphplan home site

Overview

Sensory Graphplan (SGP) is a sound, complete planner based on Blum and Furst's Graphplan. SGP includes support for conditional effects, universal and existential quantification, uncertainty, sensing actions, and uses the PDDL language syntax. SGP is written in Lisp, is efficient, and is well-suited for use in the classroom or research. The most recent version of SGP is available for download at ftp.cs.washington.edu/pub/ai/sgp.tgz.

The present version of SGP is 1.0h (dated January 14th, 2000). Version 1.0's features include:

• Conditional effects handling using Factored Expansion.
• Support for sensing actions.
• Support for uncertainty in the initial state.
• Ability to generate contingent schedules.
• Dynamic variable ordering as an optimization to backward-chaining search.
• Performance and debugging instrumentation.
• Clean, documented implementation.

Version 1.0h is a bug-fix release. See the README file in the .tgz file for a full revision history.

Publications

• Corin R. Anderson, David E. Smith, and Daniel S. Weld. Conditional Effects in Graphplan. In Proceedings of AIPS '98. 1998.
• Daniel S. Weld, Corin R. Anderson, David E. Smith. Extending Graphplan to Handle Uncertainty & Sensing Actions. In Proceedings of AAAI '98. 1998