文档介绍：A Logic Based Approach to the Static Analysis
of Production Systems
Jos de Bruijn and Mart´ın Rezk
KRDB Research Center
Free University of Bozen-Bolzano, Italy
{debruijn,rezk}***@
Abstract. In this paper we present an embedding of propositional production
systems into µ-calculus, and first-order production systems into fixed-point logic,
with the aim of using these logics for the static analysis of production systems
with varying working memories. We encode properties such as termination and
confluence in these logics, and briefly discuss which ones cannot be expressed,
depending on the expressivity of the logic. We show how the embeddings can be
used for reasoning over the production system, and use known results to obtain
upper bounds for special cases. The strong correspondence between the structure
of the models of the encodings and the runs of the production systems enables
the straightforward modeling of properties of the system in the logic.
1 Introduction
Production systems (PS) are one of the oldest knowledge representation paradigms in
artificial intelligence, and are still widely used today1. Such a system consists of a set of
rules r of the form “if conditionr then actionr”, a working memory, which contains the
current state of knowledge, and a rule interpreter, which executes the rules and makes
changes in the working memory, based on the actions in the rules.
In general rule-based systems are administered and executed in a distributed envi-
ronment where the rules are interchanged using standardized rule languages, . RIF,
RuleML, SWRL. The new system obtained from adding (or removing) the interchanged
rules need to be consistent, and some properties be preserved, . termination. In this
work we address the static analysis of such production systems, which means deciding
properties like termination and confluence. We propose using logics and their reason-
ing techniques from the area of software specifi