Type
Text
Type
Dissertation
Advisor
Kifer, Michael | Stoller, Scott | Warren, David | Liu, Yanhong | Zhou, Neng-Fa.
Date
2015-12-01
Keywords
Deductive Planning, Planning, Planning Strategies, State Space Planning, STRIPS, Transaction Logic | Computer science
Department
Department of Computer Science.
Language
en_US
Source
This work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree.
Identifier
http://hdl.handle.net/11401/77270
Publisher
The Graduate School, Stony Brook University: Stony Brook, NY.
Format
application/pdf
Abstract
Automated planning has been the subject of intensive research and is at the core of several areas of AI, including intelligent agents and robotics. In this dissertation, we argue that Transaction Logic is a natural specification, experimentation, prototyping, and potentially implementation language for planning algorithms. This enables one to see further afield and thus discover better and more general solutions than using one-of-a-kind formalisms that have been traditionally dominating this field. Specifically, we take the well-known STRIPS planning strategy and show that Transaction Logic lets one specify this strategy and several of its far-reaching extensions easily and concisely. In addition, we demonstrate the power of our approach by applying it to a fundamentally different (from STRIPS) planning strategy, called GraphPlan. To summarize, this dissertation introduces a novel planning formalism based on Transaction Logic and validates it by applying it to some existing planning strategies and developing new ones as follows: (1) we propose a non-linear extension to STRIPS planning and prove its completeness; (2) we introduce fast-STRIPS- a modification of our non-linear STRIPS extension, which is also complete and yields speedups of orders of magnitude; (3) we extend STRIPS with regression analysis, which is again complete and significantly improves performance; (4) we extend STRIPS to enable it to solve planning problems with negative derived atoms; (5) to illustrate the versatility of our formalism, we applied it to GraphPlan- a planning strategy that bears no resemblance to STRIPS. | 101 pages
Recommended Citation
Basseda, Reza, "Planning with Transaction Logic" (2015). Stony Brook Theses and Dissertations Collection, 2006-2020 (closed to submissions). 3091.
https://commons.library.stonybrook.edu/stony-brook-theses-and-dissertations-collection/3091