Evolutionary Computing

Unit Director: James Marshall

The aim of this unit is to equip you with the knowledge and skills necessary to apply evolutionary computing techniques to the solution of complex optimisation and machine learning problems. The unit describes the basic theory, mechanisms and techniques of evolutionary computing. It covers genetic algorithms, genetic programming and other evolutionary computing techniques. The unit is aimed at MSc students following the Machine Learning and Data Mining theme, and fourth year undergraduates.

Timetabled Lectures and Labs

Lectures, weeks 1-10

• Friday 10:00, 1.11(MVB)
• Friday 15:00, 1.11a(MVB)

Labs

• There are no timetabled labs, but you should make time to work through the exercise sheets

Office hour

• I aim to be in my office (MVB 3.14) on Fridays from 11:00 to 12:00 to answer course-specific questions. Non-specialist help on programming and other technical matters is available from the Help Desk (MVB 3.19).

Lecture Notes

Lecture notes will be posted here as the course progresses.

Lecture 1: Introduction	slides	print
Lecture 2: Representations and Operators	slides	print
Lecture 3: Populations and Selection	slides	print
Lecture 4: Advanced Operators and Techniques	slides	print
Lecture 5: GAs for Grouping Problems	slides	print
Lecture 6: Overview of Genetic Programming	slides	print
Lecture 7: Genetic Programming Examples	slides	print
Lecture 8: No Free Lunch Theorem	paper
Lecture 9: Leftovers from No Free Lunch	slides	print
Lecture 9: Beyond No Free Lunch	paper
Lecture 10: Schema Theory	slides	print
Lecture 11: Critiques of Schema Theory	slides	print
Lecture 12: GAs as Markov Processes	slides	print
Lecture 13: GAs as Dynamical Systems	slides	print
Lecture 14: Estimation of Distribution Algorithms	slides	print
Lecture 15: Fitness Functions and Landscapes	slides	print
Lecture 16: Other Areas of Evolutionary Computing	slides	print

Exercises

Weekly exercise sheets are posted here, with solutions posted in the following week. These exercises are unassessed, but are good preparation for the exam in January. Not all weeks will have exercise sheets, to allow you to concentrate on the coursework.

Week 1	exercises	solutions
Week 2	exercises	solutions
Week 4	exercises	solutions
Week 6	exercises	solutions
Week 8	exercises	solutions

Assessment

Assessment is by coursework (50%) and exam (50%)

• GA: Genetic Algorithm coursework - 30% - submission deadline 5th December 2008
• GP: Genetic Programming coursework - 20% - submission deadline 12th December 2008
• Exam: Theory exam - 50% - January 2009

GA Coursework	description	data	data format	results (OpenOffice format)
GP Coursework	description	data	results (OpenOffice format)

Recommended Reading

Lecture notes are provided during the course, but the following are also excellent references.

An errata sheet for Reeves and Rowe is available.

• Reeves, C. R. and Rowe, J. E. Genetic Algorithms - Principles and Perspectives: A Guide to GA Theory. 2003. Kluwer Academic Publishers. ISBN: 1402072406 Price: £88.50 Recommended
  
• Goldberg, D. E. Genetic Algorithms in Search, Optimization and Machine Learning. Addison-Wesley Publishing Company. 1989. ISBN: 0201157675 Price: £56.99 Background
  
• Mitchell, M. An Introduction to Genetic Algorithms 1998. MIT Press ISBN: 0262631857 Price: £21.95 Background
  

The recommended and background texts are in the library.