Stephanie Forrest
Office: Farris Engineering Center 355E
Phone: 277-7104 (during office hours only)
Email: forrest@cs.unm.edu
Office Hours: Tue, Thu 11-1 (or by appointment)

An Introduction to Genetic Algorithms by Melanie Mitchell.

Genetic Algorithms in Search, Optimization, and Machine Learning by David Goldberg (optional).

Selected Readings.

Evolution by natural selection is one of the most compelling themes of modern science, and it has revolutionized the way we think about biological systems. In this course, we will study a form of evolution, called genetic algorithms, that takes place in a computer. In genetic algorithms, selection operates on populations of individuals stored in the computer's memory, and over time the functionality of these individuals evolves in much the same way that natural populations of individuals evolve. While the computational setting is highly simplified when compared with the natural world, genetic algorithms are capable of evolving surprisingly complex and interesting structures. These structures (individuals) may represent solutions to problems (e.g., parameter settings), strategies for playing games, visual images, or even computer programs.

This is in part a discussion class, and there will be a large number of assigned readings. Students are expected to participate in class discussions and 25% of the assigned grade for the class will depend on class discussion. You should plan on reading carefully (2 - 3 times) approximately two scientific papers per week along with the assigned textbook reading. In addition, there will be a number of assignments throughout the semester, including a final project (with writeup) and (depending on the class size) an oral presentation.

I. Course Introduction (2 weeks)

Topics:

Evolution, genetics, overview of genetic algorithms

Readings:

Darwin Origin of Species (Ch. 1 - 3)
Mitchell (Ch. 1)
Forrest (Science, 1993)
[Goldberg (Ch. 1 - 2)]**
Back and Schwefel (ECJ, 1993)

II. Implementation Techniques (3 weeks)

Topics:

Data structures and representation, selection, fitness scaling, crossover (1 pt, 2 pt, uniform), elitism, hybrid GAs, genetic programming, etc.

Readings:

Mitchell (Ch. 5)
[Goldberg (Ch. 3,5)]
Ono and Kobayashi (ICGA 97)
Miller and Goldberg (ECJ, 1996)
Thierens (ICGA 97)
Sarma and DeJong (ICGA 97)
Eshelman et al. (ICGA 89)
Booker (FOGA 2)
Eshelman et al. (ICGA 97)
Mitchell (pp. 35-44)

III. Using Genetic Algorithms to Solve Problems (3 weeks)

Readings:

Mitchell (pp. 44-83)
[Goldberg (Ch. 4)]
Hillis (Physica D, 1990)
Belew and Rosen (ECJ, 1997)
Neubauer (ICGA 97)
Rogers (ICGA 97)
Oliver et al. (ICGA 87), Starkweather et al. (ICGA 91)
Sims (Computer Graphics, 1991)

III. Theory (3 weeks)

Topics:

Schema anlaysis, deception, schema variance, landscape analysis, population genetics models, idealised GAs

Readings:

Mitchell (Ch. 4)
[Goldberg (Ch. 2)]
Forrest and Mitchell (Machine Learning, 1993)
Nimwegen et al. (SFI, 1997)
Christiansen and Feldman (1997)

V. Using Genetic Algorithms as Models (3 weeks)

Readings:

Mitchell (Ch. 3)
Holland et al. (Induction, 1986; selected readings)
Holland (Hidden Order, 1995; selected readings)
Forrest et al. (Immune system modeling)

VI. Student Presentations (2 weeks)