Backward chaining inference as a database stored procedure – the experiments on real-world knowledge bases

Tomasz Xie¸Ski,Roman Simiński

doi:10.1080/24751839.2018.1479931

Tomasz Xie¸Ski, Roman Simiński

Open Access

https://doi.org/10.1080/24751839.2018.1479931

Copy DOI

Abstract

ABSTRACTIn this work, two approaches of backward chaining inference implementation were compared. The first approach uses a classical, goal-driven inference running on the client device – the algorithm implemented within the KBExpertLib library was used. Inference was performed on a rule base buffered in memory structures. The second approach involves implementing inference as a stored procedure, run in the environment of the database server – an original, previously not published algorithm was introduced. Experiments were conducted on real-world knowledge bases with a relatively large number of rules. Experiments were prepared so that one could evaluate the pessimistic complexity of the inference algorithm. This work also includes a detailed description of the classical backward inference algorithm – the outline of the algorithm is presented as a block diagram and in the form of pseudo-code. Moreover, a recursive version of backward chaining is discussed.

Highlights

Knowledge-based systems are still popular and practically used tools for solving ill-structured problems
We present the formal description of a knowledge base, the backward chaining inference algorithm, a brief review of software tools related to this work and a short description of our own software implementation
For each rule r [ R we define the following functions: concl(r) – the value of this function is the conclusion literal of rule r: concl(r) = c; cond(r) – the value of this function is the set of conditional literals of rule r: cond(r) = {p1, p2, . . . pm}, literals(r) – the value of this function is the set of all literals of rule r: literals(r) = cond(r) < {concl(r)}

Summary

Introduction

Knowledge-based systems are still popular and practically used tools for solving ill-structured problems. The number of applications which utilize rule bases and methods of inference grows, but the number of tools for building knowledge-based systems increases much more slowly (Sajja & Akerkar, 2010). We present the formal description of a knowledge base, the backward chaining inference algorithm, a brief review of software tools related to this work and a short description of our own software implementation. The following formal description of a knowledge base is assumed in this work: a knowledge base is a pair KB = (R,F) where R is a non-empty finite set of rules and F is a finite set of facts. Each fireable rule can be activated, the conclusion of activated rule is added to facts set F –activate(r) : F = F < {concl(r)}

Objectives

Methods

Results