This project is supported in part by the ECSEL/ENSYM Excellence Center in Computer Science and Systems Engineering at Linköping University; TFR, the Swedish Council for Engineering Sciences; and the Wallenberg Foundation.
A description of research activities in this project are structured as follows:
This project focuses on the development of research and application tools to support the modeling and reasoning about dynamic processes which arise due to interactions between physical or software artifacts and the complex dynamic environment in which they are situated, or among several physical artifacts situated in similar environments. The environments of interest may be characterized by the inability to provide complete representations of their dynamics due to their complexity.
The project includes development of temporal logics for reasoning about action, change and processes; the development of inference techniques for use in robotics systems; the development of visuallization tools to support this type of research; the application of temporal logics to planning, prediction, explanation and diagnosis; and development of tools to support the structuring of large narrative applications using object-oriented techniques. The rest of this page describes our activities both past and present in this area.
One of the primary reasoning tools used to specify both artifacts, environments and interactions is a narrative-based formalism called TAL which includes a family of Temporal Action Logics suitable for modeling in the context of incompletely specified domains. TAL has a long history inspired by the work of Erik Sandewall with Features and Fluents in 1994. Many extensions to the original version have been developed during the period 1994-2001, several are described on this page. A selected list of publications can be found at the end of this page.
One current focus of the project is the development and specification of a tool for reasoning about temporal narratives in TAL which combines the power of both symbolic and visual modalities. Narratives may be represented in sequential text form and their models depicted in diagrammatic form in terms of time lines. A querying tool is included for reasoning about specific narratives. The tool, called VTAL, is accessible via the WWW in a number of forms and has been since 1997.
Two major and useful extensions to the original version of TAL have been developed
A number of other extensions and applications based on these have been developed such as delayed effects of actions. Descriptions of these may be found in the articles listed below.
TAL-OBJ is a recent development used to structure the modelingof large applications where narratives and their expansions can involve 100's of logical formulas. Object-Oriented structuring techniques have been developed for structuring large axiomatizations. This eases the design and debugging of specifications in complex application domains and has the potential for contributing to the optimization of inference techniques by taking advantage of the module representation of the narrative axioms.
TALplanner is recent development and has its own Home Page. TALplanner is a logic-based forward chaining planner based on the use of TAL. It recently won the domain-independent planning competition at the Artificial Intelligence Planning and Scheduling Competition, AIPS' 2000. It also won first place in the Schindler Elevator Control planning competition.
A relatively up-to-date description and overview of the TAL family of logics:
Temporal Action Logics: Language Specification and Tutorial.
Doherty, P., Gustafsson, J., Karlsson, L., and Kvarnström, J.
Electronic Transactions on Artificial Intelligence, Volume 2, Issue 3-4, pages 273-306, 1998.
A proposal to deal with the qualification problem by using dependency constraints:
Tackling the Qualification Problem using Dependency
Kvarnström, J. and Doherty, P.
Computational Intelligence, Volume 16, Number 2, pages 169-209, May 2000.
An extension to TAL which provides a means of modeling delayed effects of actions:
Effects of Actions.
Karlsson, L., Gustafsson, J., and Doherty, P.
Proceedings of the 13th European Conference on Artificial Intelligence, Aug. 23-28, Brighton, pages 542-546 (ECAI-98), 1998.
An extension to TAL which provides a means for modeling concurrent actions and their interactions:
Reasoning about Concurrent Interaction.
Karlsson, L. and Gustafsson, J.
Journal of Logic and Computation, Volume 9, Issue 5, Oct 1999.
An earlier version called Reasoning about Action in a Multi-agent Environment is available at Linköping University Electronic Press (1997).
An extension to TAL which introduced the idea of dependency constraints and provides a basis for dealing with the ramification problem:
Occlusion in Specifying the Indirect Effects of Actions.
Gustafsson, J. and Doherty, P.
Proceedings of the 5th International Conference on the Principles of Knowledge Representation in Reasoning (KR'96), 1996.
The original node article from which the remaining work has been developed. This provided a syntactic characterization of a definition of preferential entailment in Sandewall's Features and Fluents work in terms of Circumscription and showed that the 2nd-order characterization could be reduced to a 1st-order equivalent:
Reasoning about Action and Change using Occlusion.
Doherty, P. (1994).
Proceedings of the 11th European Conference on Artificial Intelligence (ECAI-94), Aug. 8-12, Amsterdam, pages 401-405, 1994.
An out-of-date specification of the original formalism during the transition from simple PMON to TAL. Contains some unpublished material:
For information about Features and Fluents and the original definition of PMON preferential entailment, see Sandewall[FF].