Suspect identification using fuzzy Petri

Dissertant

Laftah, Husayn Atiyyah

University

University of Technology

Faculty

-

Department

Computer Sciences Department

University Country

Iraq

Degree

Ph.D.

Degree Date

2004

English Abstract

In this work, the structural properties of Petri nets and fuzzy logic are applied for handling the problems of fuzzy reasoning in expert systems, in the presence of imprecision and inconsistency of database, and uncertainty of a knowledge base. A fuzzy Petri net model (FPN) is built from a given set of suspects, fuzzy production rules of a rule-based system, and a database.

An algorithm, which combines pivot and invariant methods, is used for the analysis of FPN model, in order to detect and eliminate the cyclic behavior which occurs in the fuzzy Petri net.

Second algorithm is developed to detect and eliminate the contradiction which occurs in the fuzzy Petri net model.

Third algorithm is also developed to compute the steady state of fuzzy beliefs at time (t) from its initial value.

These steady state values of nodes at terminals are used to identify the culprit.

Fourth algorithm is used to find the useful part of the network for generating an evidential explanation for the culprit.

The worst time complexity of computing the steady state and explanation are Q (m), 0 (an), where m is the number of transitions, n is the number of places and a is the number of axioms. The package SIFP (Suspect Identification using Fuzzy Petri Net) is designed and implemented using visual Basic6 as programming language, and the minimum operation requirement is Pentium computer.

Main Subjects

Information Technology and Computer Science

Topics

• Fuzzy logic

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-305585