MKM 502E - Soft Computing

Course Objectives

1)To teach students ,the fundamentals of artificial neural networks, training algorithms, and to show them how neural networks are used in control problems.
2) To teach the basics of fuzzy logic, fuzzy logic inference mechanism, and how fuzzy logic is used in controller design.
3) To inform students about derivative-free global optimization methods, with an emphasis on genetic algorithms.
4) To make students implement the methods they have learned in this course to a mechatronic system in the framework of a term project and to make both written and oral presention of their work.

Course Description

Learning and types of learning, optimization techniques, introduction to neural networks, single layer perceptrons, multi-layer perceptrons, backpropagation algorithm, implementation of neural networks on control problems, fuzzy sets, operations on fuzzy sets, fuzzy relation and composition, fuzzy inference systems, fuzzy controllers, adaptive neuro-fuzzy inference systems (ANFIS), radial basis functions (RBF), genetic algorithms, other derivative-free global optimization methods