ECE 657 - Computational Intelligence / Intelligent Systems Design
Instructor
Professor
Fakhri
Karray
Telephone
extension
35584
Room
E5
5029
email
Course outline
Conventional approaches for tackling complex systems are usually modeled/implemented under the assumption of a good understanding of the process (system) dynamics/functionalities and its operating environment. These techniques fail, however, to provide satisfactory results when applied to ill-defined systems, for which analytical and experimental models are not easily obtained. These same systems may also operate within unpredictable and possibly noisy/uncertain environment. Recent developments in the area of intelligent systems and soft computing have presented powerful alternatives for dealing with the behavior of this class of systems. This course outlines fundamentals of soft computing based design approaches using such tools as approximate reasoning, fuzzy inferencing, neural networks, evolutionary algorithms, and neuro-fuzzy systems. Fundamentals and advances on these procedures are outlined along with their potential applications to various real world applications in virtually most fields of engineering including pattern recognition, system planning, data analysis, classification, power generation, intelligent transportation, systems and control, intelligent mechatronics, optimization, communication, robotics and manufacturing, to name a few.
Lecture time
Tuesdays and Thursdays from 1:00 pm to 2:20 pm, Room E5 5106
Main text
Material taught in this course and assignments provided are from the following textbook:
- F. Karray and C. de Silva, Soft Computing and Intelligent Systems Design, Addison Wesley Publishing, Pearson Education, August 2004
Other suggested texts
- M. Negnevitsky, Artificial Intelligence, A Guide to Intelligent Systems, Pearson Publishing, 2006
- C. T. Lin and C.S. Lee, Neural Fuzzy Systems, Prentice Hall Publishing, 1995
- J. Jang, C. Sun, and E. Mizutani, NeuroFuzzy and Soft Computing, Prentice Hall Publishing, 1997
Course scope
The course is useful for graduate students in virtually all areas of engineering, particularly for those dealing with complex systems at the sort or hard levels. A background in two or more of the following areas should be useful: fuzzy logic, neural networks, system's optimization, dynamic systems, nonlinear mapping, calculus of variation, differential calculus, statistical analysis, advanced algebra.
Course material and online resources
All course material, including slides, notes, assignments, exams are posted on the course page on Learn. Exams and final projects are uploaded on the course web page on Learn.
Course requirement
Assignments (0%): Assignments will be provided on a biweekly to tri-weekly basis. Students are highly encouraged to solve them.
Midterm (25%): A midterm will be given around last week of June.
Journal Paper Analysis and Presentation (20%): The student is required as part of the course workload to analyze, and possibly synthesize results of one or more journals publication in one of the areas related to the course.
Final Exam (55%): As the final exam requirements, the student is to work on final project illustrating and making use on the tools taught during the term.
Tentative course outline
The main sections of this course are given as follows:
- Introduction
- Approximate Reasoning
- Fuzzy Inferencing and Intelligent Systems
- Connectionist Modelling, Classification and Support Vector Machines
- Evolutionary Techniques
-
Hybrid
System
Design
- Neuro-Fuzzy Systems
- Gentic-Neuro Systems
- Applications
Library material
A large set of relevant of journals and texts related to the subject are available in the library or online, including:
- IEEE Transactions on Fuzzy Systems
- IEEE Transactions on Neural Networks
- IEEE Transactions on Evolutionary Computation Fuzzy Sets and Systems