JPH03211625A - Rule set shift method for expert system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION A. Field of Industrial Application The present invention relates to an expert system, and particularly to a method for migrating a rule set in a proxy rule expression.

B0 Summary of the Invention The present invention provides an expert system that performs inference by transitioning between grouped rule sets, in which rule set transition order information is created in a first rule set, and the information is passed to arguments and 1. This ensures that rulesets can be migrated reliably and easily by allowing other rulesets to be migrated sequentially.

C. Prior Art The Exverton stem is constructed as shown in FIG. The knowledge of experts in the field is shared by experts.
”:) organizes the information and stores it in the knowledge base 1, and the inference system 2 obtains answers to questions from the user by inference using the knowledge in the knowledge base 1.

The inference system 2 has a description language 3 such as L TSP and PROLOG, uses a knowledge base editor 4 and a man-machine interface of a developer 5 to store knowledge in the knowledge base 1, and uses a man-machine interface for inference. While securing data such as knowledge base areas, rules, facts, etc. necessary for the work process in the working memory 6, the inference engine 7 performs explanatory reasoning.

In addition, an explanation function 8 is provided on the man-machine interface with the user to accept questions and diagrams and display inference procedures.

Here, the knowledge expression of knowledge base 1 includes IF<condition part>
Production rule expressions in the form of "THF":N<conclusion part> and frame expressions in which abstract concepts and entities are each expressed as frames in a hierarchical network are known, and production systems and frame systems that make inferences based on these knowledge, as well as A hybrid system that combines the two is known.

Also, known inference methods include forward inference leading from the condition part of a rule to the conclusion part, and backward inference leading from the conclusion part of the rule to the condition part.

In addition to such knowledge rules and reasoning methods, efficiency, speed, and functionality of reasoning can be improved by various methods such as conflict resolution methods, working memory manipulation methods, meta-rule creation methods, and belief change introduction.

Here, a conventional method of knowledge rules and inference using progukun9 rule expressions will be explained. The rules in Knowledge Base 1 are grouped by related rules into rule sets.
Inference efficiency is increased by performing inference on a rule set basis. FIG. 3 shows an inference method using a ruleset in a schematic diagram of ruleset migration. Each of the rule sets A to D has a meta rule for controlling the rules within the rule set, and the name of the rule set to be used next is written in the meta rule. With this meta-rule description, for example, after inference is made about each rule of rule set A, the process moves to rule set B, inference is made about the rules in B, and when this inference is completed, the process moves to the next rule set C.

D Problems to be Solved by the Invention In the conventional ruleset transition method, when transitioning to ruleset B is described in the meta-rule of ruleset D as shown in the third diagram, the transition to ruleset B is performed again after transitioning to ruleset B. The process repeats the transition from rule set 01 to rule set D, sometimes resulting in an endless loop between specific rule sets.

Also, when adding another ruleset to an existing ruleset, it is necessary to modify the meta-rules of the ruleset that is applied before the added ruleset.
Similarly, when deleting or changing the order of a rule set, the meta-rules must be modified, and adding, deleting, or changing the order of rule sets is time-consuming and prone to mistakes.

The purpose of the invention is to ensure the migration of the Lurse throat.

The purpose of the present invention is to provide an easy migration method.

E. Means and Effects for Solving the Problems In order to achieve the above object, the present invention provides a plurality of rule sets in which rules of production rule expression are grouped, and an inference engine migrates between the plurality of rule sets. In the system that performs inference using a first rule set, a first rule set is provided to create a sequence of rule sets that becomes the transition order information of the rule sets among the rule sets, and rules to be applied in the inference based on this first rule set are provided. an argument that makes the transition to the first rule set in the sequence of rulesets when the rule disappears or a certain condition is met, and sets the second rule set in the sequence of rulesets as the next transition destination rule theft. Create a sequence of rulesets by inference using the first ruleset, migrate to the first ruleset in this sequence, and make the second ruleset the next migration destination ruleset. Repeat migration for each ruleset by passing arguments.

F. Embodiment FIG. 1 is a schematic diagram of rule set transition showing an embodiment of the present invention. Rule set A includes rules for creating an arrangement of rulesets, and the arrangement of rulesets is created from the inference result using this ruleset node A. This arrangement of rulesets becomes order information of rulesets used for inference. The illustration shows a case where 'RCr)C] is created as a sequence of rule sets.

When there are no more rules to apply in the inference using rule set A, or when certain conditions are met, migration is made with rule set B, which is the first in the list of rulesets created in rule set A, as the destination ruleset. and the second and subsequent [Cr)C of the order information of the ruleset as an argument.
Pass I to ruleset B. Using this rule set B, inference is performed until there are no more rules to apply or a certain condition is met, and after this, the second and subsequent arguments [T) CI
is passed to ruleset C as a new argument. Similarly, a new argument is passed from ruleset C to the next ruleset D (7) This transition is repeated until there is no more argument data. In the example shown, ruleset A4R-4C-D→
Migration of the ruleset=t is performed in the order of C.

Therefore, the migration order of the rulesets is determined by creating the order of the rulesets in the ruleset A, and even when a certain ruleset is migrated again, an endless loop will not occur. Additionally, to add or delete rulesets or change the migration order, all you need to do is add or delete the ruleset name in ruleset A, which creates the order of rulesets, and there is no need to change the meta-rules of other rulesets. While making changes to the ruleset optional.

Effects of the G8 Invention As described above, according to the present invention, a sequence of rulesets is created by inference based on the first ruleset, and an argument that specifies that the second ruleset in this sequence is the destination ruleset is set. Since the migration of ruleset nodes is performed by passing, the occurrence of an eternal loop in the conventional method is eliminated, and the creation of the order of rulesets by the first ruleset node is changed to add, delete, and change the order of rulesets nodes. It is easy to do with a chisel and has the effect of reducing mistakes.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram of a ruleset transition showing an embodiment of the present invention, FIG. 2 is a basic configuration diagram of an Exverton stem, and FIG. 3 is a schematic diagram of a conventional ruleset transition. 1...Knowledge base, 7...Inference engine. 2 people outside Figure 2

Claims (1)

[Claims] (1) In an expert system in which a group of production rule expression rules are grouped into multiple rule sets, and an inference engine performs inference by transitioning between the multiple rule sets, information on the transition order of the rule sets among the rule sets. A first ruleset is provided that creates a sequence of rulesets that becomes At the same time as moving to the ruleset, the sequence 2 of the ruleset
An expert system ruleset migration method characterized by passing an argument that specifies the next migration destination ruleset as the next migration destination ruleset.