Modeling hybrid domains using process description language

Date

2004-12

Journal Title

Journal ISSN

Volume Title

Publisher

Texas Tech University

Abstract

Researchers in the field of knowledge representation and logic programming are constantly trying to come up with better ways to represent knowledge. One of the recent attempts is to model dynamic domains. A dynamic domain consists of actions that are capable of changing the properties of objects in the domain, for example the blocks world domain. Such domains can be modeled by action theories - collection of statements in so called action languages specifically designed for this purpose. In this thesis we extend this work to allow for continuous processes properties of objects that change continuously with time. For example the height of a freely falling object. In order to do this we adopt an action language/logic programming approach.

A new action language called process description language is introduced that will be useful to model systems that exhibit both continuous and discrete behavior (also called hybrid systems). An example of a hybrid domain is the domain consisting of a freely falling object. A freely falling object is in the state of falling, which is a discrete property that can be changed only by actions (also called fluent) while its height is a continuous process.

The syntax, semantics, and translation of the statements of the language into rules of a logic program will be discussed. Examples of domains that can represented in this language will be given. In addition, some planning and diagnostic problems will be discussed. Finally, the language will be compared with other languages used for similar purposes.

Description

Keywords

Logic programming languages, Knowledge representation (Information theory), Computer simulation

Citation