|
Soft
Computing
CSE 513 Fall 2005 (4 cr) Tuesday
& Thursday, from 5:30 to 7:17 pm, Room ---
Instructor
Djamel Bouchaffra, PhD
Assistant Professor of Computer
Science
Office: 131 Dodge Hall
Phone: (248) 370-2242
Fax: (248) 370-4625
Office Hours: Tuesday
& Thursday from 3 to 5 pm
Description
We
present in this course a detailed introduction to the theory and terminology of
fuzzy disciplines and adaptive environments. We explain the concepts of fuzzy
sets, fuzzy rules, fuzzy inference systems and fuzzy reasoning. We also provide
an overview of model (or system) identification and optimization techniques
which are used in neural-fuzzy systems. We will tackle the adaptive neural
networks, supervised learning, learning from reinforcement and the unsupervised
learning.
Least-squares
methods will be introduced for model identification. We introduce nonlinear
optimization using derivative-based techniques such as gradient and derivative
free techniques such as genetic algorithm, simulated annealing, downhill simplex
method and random search.
Some
practical problems that involve fuzzy and adaptive systems will be solved by the
students during this course.
Prerequisites:
CSE 501, CSE 504; Basic knowledge of elementary calculus
and linear algebra; Knowledge
of MATLAB is highly recommended.
|
