GCO4015

Evolutionary and neural computing

B Nath and M Mohammadian

6 points - 4 hours per week - Second semester - Gippsland, distance - Prerequisites: GCO2802, GCO3815 or similar - Prohibitions: Nil

Objectives Students completing the subject will be aware of the nature of intelligent systems and their applications; have detailed knowledge of some common neural network architectures; be able to manipulate genetic operators and design neural nets; understand techniques for numerical optimisation; be able to apply the techniques studied in the subject to practical problems.

Synopsis Introduction to intelligent systems. Concepts of genetic operators such as crossover and mutation. Fitness functions, scaling and sampling in GAs. Numerical optimisation using GAs. Applications to scheduling problems and machine learning. Basic concepts of neural computing. Introduction to some types of neural networks: feedforward neural nets, autoassociative nets, self-organising nets and fuzzy and neurofuzzy systems, and applications to which they are suited. Design of neural computing applications and optimisation using an iterative approach. Basic concepts of simulated annealing and applications.

Assessment Assignments: 60% - Examination (2 hours): 40%

Prescribed texts

To be advised

Back to the 1999 Information Technology Handbooks