Dismiss
Surety is performing system maintenance this weekend. Electronic date stamps on new Prior Art Database disclosures may be delayed.
Browse Prior Art Database

# Fast Fuzzy Inference Engine

IP.com Disclosure Number: IPCOM000122699D
Original Publication Date: 1991-Dec-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 3 page(s) / 77K

IBM

## Related People

Ishikawa, S: AUTHOR

## Abstract

Disclosed is a method for a fast fuzzy inference engine that includes a fuzzy-rule searching method using a hash function. A fuzzy-rule has fuzzy variables in input parameter conditions.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 72% of the total text.

Fast Fuzzy Inference Engine

Disclosed is a method for a fast fuzzy inference engine
that includes a fuzzy-rule searching method using a hash function. A
fuzzy-rule has fuzzy variables in input parameter conditions.

A conventional procedure for a fuzzy inference engine is as
follows (Fig. 1):
1) Fuzzy variable calculation of input parameter.
2) Rule fitness (Wi) calculation.
3) Accumulation of output fuzzy variables.
4) Output calculation (center of gravity of output fuzzy variables).

In step 2), the conventional method searches all of the rules
in order to calculate a fitness of the rule. However, if one of the
input fuzzy variables of the rule is zero, the fitness of the rule
will be zero. It is not necessary to evaluate the fitness of the
rule.

In the conventional method, they have tried to improve the
calculation speed of the membership value of the fuzzy variables by
using the calculation table. However, we have found no improvement in
the rule-fitness calculation.

Fuzzy inference has become popular in various areas of
intelligent control systems. This idea should improve the fuzzy
inference performance, especially in the case of a large number of
rules.
Algorithm

In the calculation of the fitness of a fuzzy-rule, if one of
the input fuzzy variables does not fit the condition of the rule, the
rule of the fitness will be zero. When fuzzy inference is done, the
conditions of the fuzzy-rules are not changed....