Department of Computer Science
Washington & Lee University
CS397 (Fall 2002): Genetic Algorithms
Course instructor: Prof. Simon Levy (Parmly 407B, x8419, levys@wlu.edu).
Office hours: MWF 2-4 and by appointment.
This page: http://simondlevy.academic.wlu.edu/courses/csci-397-genetic-algorithms-f02/
Course Requirements
- Prereqs: Programming competence (Java, C, or C++); a little theory (inductive proofs)
- Homework (“thought exercises” and coding)
- Final programming project (individual or team)
- Class participation: This is a seminar, and a goal will be to have everyone present a paper or unassigned section of the book, as well as a final project summary, if we have time.
- Each of these three (homework, project, participation) will count for approximately 1/3 of your final grade. This policy is flexible; for example, if you have trouble with some of the problem sets but do an outstanding final project, the project grade will count more.
PROBLEM SETS
Guidelines – please read them carefully and be sure that you understand them.
- Problem Set 1 (due Monday 23 September). Postscript PDF
Java Roulette wheel class: Class file | Documentation | Source code
Double-Precision Floating-Point Arrays class: Class file | Documentation | Source code - Problem Set 2 (due Friday 11 October). Postscript | PDF
BoolTree class: Zip file | Documentation | Source code
New Java Roulette wheel class: Class file | Documentation | Source code
New Double-Precision Floating-Point Arrays class: Zip file | Documentation
FINAL PROJECT: Proposal Due Monday 16 October
This project can be as simple as replicating published results; for example, implementing Hillis’s sorting networks GA, or something equally non-trivial. Or, more likely, it will be a research problem from your own area of interest (math, geology, programming languages) that you’d like to try and solve with a GA. The actual project should take about a month of work, and be summarized in a 5-10 page writeup, with references, that you will present to the class at the end of the term.
LECTURE NOTES
- 06 September 2002 Postscript PDF (Intro, Course Outline)
- 09 September 2002 Postscript PDF (Crossover/Mutation, Fitness Landscapes, No Free Lunch)
- 11 September 2002 Postscript PDF (Prisoner’s Dilemma, Sorting Networks)
- 13 September 2002 Postscript PDF (Schemas, Building Blocks)
- 15 September 2002 Postscript PDF (Genetic Programming)
- 20 September 2002 Postscript PDF (Cellular Automata)
- 23 September 2002 Postscript PDF (Intro. to Neural Networks)
- 25 September 2002 Postscript PDF (Evolving Neural Networks)
- 27 September 2002 Postscript PDF (Evolving a Learning Rule)
- 30 September 2002 Postscript PDF (The Baldwin Effect)
- 09 October 2002 Postscript PDF (Theoretical Foundations)
- 14 October 2002 Postscript PDF (When to use a GA; Encodings)
- 16 October 2002 Postscript PDF (Inversion; Hot Spots)
- 18 October 2002 Postscript PDF (The Messy GA; Selection Methods)
October – November 2002 Postscript PDF (Watson’s SEAM Algorithm) - 22 November 2002 Postscript PDF (MBOA: Multi-objective Bayesian Optimization Algorithm)
LINKS
- AI Helps Gamers Keep on Playing : Using a GA to learn a human player’s strategy (Check out the nerd in the byline!)
- The DEMO Lab : My old stomping ground! Lots of cool GA-related papers, interactive games, animations, etc.