Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Backward Chaining at Compile Time

IP.com Disclosure Number: IPCOM000037538D
Original Publication Date: 1989-Mar-01
Included in the Prior Art Database: 2005-Jan-29
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Gallivan, HW: AUTHOR [+4]

Abstract

This article discloses a technique for performing backward chaining reasoning with a forward chaining inferencing mechanism for the development of knowledge based systems applications. In inference engines, there are primarly two methods of reasoning: backward chaining, which is a goal-driven mechanism, and forward chaining, which is fact-driven.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 78% of the total text.

Page 1 of 1

Backward Chaining at Compile Time

This article discloses a technique for performing backward chaining reasoning with a forward chaining inferencing mechanism for the development of knowledge based systems applications. In inference engines, there are primarly two methods of reasoning: backward chaining, which is a goal-driven mechanism, and forward chaining, which is fact-driven.

In some applications, such as diagnostic applications, it is desirable to utilize a backward chaining strategy. In other applications, a forward chaining strategy is more desirable. In order to provide flexible tools for the development of a wide range of knowledge based applications when forward chaining is used as the primary inferencing mechanism, backward chaining can be implemented by actively soliciting the facts needed to satisfy a given goal. By adding a mechanism to select goals and to request the facts required by those goals, backward chaining can be simulated. When the facts are acquired, they are entered into a network in the same manner as forward chaining data and propagated to the goals. Augmentations to the forward chaining reasoning network can be used to maintain the set of goals and facts needed in order to prevent unnecessary or extraneous goals from being pursued.

To improve the performance of this process, the goals and data used by backward chaining can be analyzed at compilation or preprocessing time (prior to execution of the knowledge base). This analysis resul...