Combining action and process specification for high-level agent control

Información del proyecto

Identificador del acuerdo de subvención: FMBI961574

Fecha de inicio

Fecha de finalización

Financiado con arreglo a

• Specific research and technological development programme in the field of the training and mobility of researchers, 1994-1998

Coste total

Sin datos

Aportación de la UE

Sin datos

Coordinado por

Aristotle University of Thessaloniki
Greece

Objetivo

Research objectives and content With my doctoral work I believe to have proposed advancements to the theory of actions and change in the field of knowledge representation: i) the definition of specification languages that capture increasingly-complex aspects of reasoning about actions, in particular the treatment of actions actually occurred in the domain and hypothetical actions was reconciled; ii)the discussion and formalization of causal explanations. These specification languages may be seen as logics of actions, in the sense that they have an ad hoc model theory and a notion of entailment. The emphasis, however, was on the computer science aspects, e.g. the ability of a formalization to assimilate a change in the domain by simple addition of facts and the ability to describe actions intended to be performed by an autonomous agent. Yet these languages have a rather limited expressive power. Two notable limitations are: i) only fluents with boolean values are allowed, and therefore representing continuously changing values, e.g. position, becomes impractical; ii) there is no direct way to specify the performance of a repetition of actions so as to achieve/maintain a certain state. The latter, which is instead unproblematic in process specification languages, has not yet suggested a cross-fertilization of ideas between knowledge representation and concurrency. In my stay at Queen Mary and Westfield college I intend to study solutions to the two issues above. This undertake involves also a change in methodology, as I am going to follow a more a experimental track, i.e. to draw a research agenda which is based on experiences in specifying small-scale realistic agents. Training content (objective, benefit and expected impact) The time allocated for the project seems sufficient to bring us to the definition of the extended language, implementation of a prototype interpreter and experimental application to governing of a wandering robot that should already be available at Queen Mary and Wetsfield.

Ámbito científico (EuroSciVoc)

• ciencias naturales › matemáticas › matemáticas puras › matemáticas discretas › lógica matemática
• ciencias naturales › informática y ciencias de la información › ingeniería del conocimiento

Programa(s)

• FP4-TMR - Specific research and technological development programme in the field of the training and mobility of researchers, 1994-1998

Tema(s)

• 0302 - Post-doctoral research training grants

Convocatoria de propuestas

Datos no disponibles

Régimen de financiación

Datos no disponibles

Coordinador

Aristotle University of Thessaloniki

• Aportación de la UE: Sin datos
• Dirección: Analytical Chemistry Lab ; 54006 Thessaloniki ; Grecia
• Coste total: Sin datos