First Things First This is 4003-590-02 / 4005-756-02 Welcome to Genetic Algorithms (Genetic Algorithms) I am Joe Geigel…your host! Plan for this afternoon First some thanks Logistics To Prof Butler for filling in for me. What is this course about? To WGBH for making that engrossing video on evolution. Requirements and Deliverables Time for attendance. Logistics Logistics Course Web Site: Course Web Site: http://www.cs.rit.edu/~jmg/geneticAlgorithms http://www.cs.rit.edu/~jmg/geneticAlgorithms Contact: Everything you need to know office hours: MW 2-4 (or by appt) Syllabus Office: 70 (GCCIS) Rm 3527 Assignments e-mail: jmg@cs.rit.edu Schedules phone: 475-2051 Reading List Diary Slides: Will be available on Web site. 1
Logistics Logistics Course Web Site: Your RIT e-mail http://www.cs.rit.edu/~jmg/geneticAlgorithms Be sure that it is forwarded correctly. Will be the authoritative source for the course. myCourses E.g. SCHEDULE has already been Email changed. dropboxes Logistics What is this course about? Textbook Evolutionary Computation by DeJong Selected papers from GA literature (on Web site) Supplement lectures Not necessary, but nice to have as references. Top 5 misconceptions about this course. Evolutionary Algorithms Only cover GAs An EA uses some mechanisms inspired by biological 1. evolution: reproduction, mutation, recombination, Will have an awfully hard final exam 2. natural selection and survival of the fittest. Hey, aren’t you Prof Anderson? 3. Candidate solutions to the optimization problem play Will cover an awful lot of theory (leading to 4. the role of individuals in a population, and the cost an awful final exam) function determines the environment within which Long PBS documentaries will be shown each the solutions "live". 5. class. Evolution of the population then takes place after the repeated application of the above operators. 2
Course objectives Getting back to the video Upon completion of this course, students will Things that you unexpectedly got from the be able to: video: Explain the of the principles underlying Monkeys are capable of learning the concept of 0 Evolutionary Computation in general and Genetic Charles Darwin's brother is one of the best roles Algorithms in particular. I've ever seen on a PBS documentary. Apply Evolutionary Computation Methods to find Darwin married his cousin solutions to complex problems Analyze and experiment with parameter choices in finches can be blown really really far by the wind. the use of Evolutionary Computation There were some really big rodents at one point Summarize current research in Genetic Algorithms and Evolutionary Computing (GRAD Only) Thing I hoped you’d get from Thing I hoped you’d get from the video the video 1. Individuals evolve over generations. 2. Evolution is guided by fitness in a given environment. Thing I hoped you’d get from Thing I hoped you’d get from the video the video 3. Individuals are the “product” of their 4. An individuals traits can undergo parents. random mutation 3
Thing I hoped you’d get from the video Getting back to the video 6. Nature has no apparent 5. Genotype vs. phenotype goal (except to create individuals suitable to a given environment). Evolutionary Algorithms Evolution and Computing An EA uses some mechanisms inspired by biological evolution: Your “world” is the context of a given reproduction, mutation, recombination, natural selection and problem survival of the fittest. Candidate solutions to the optimization problem play the role of individuals in a population, and the cost function determines the environment within which the solutions "live". Evolution of the population then takes place after the repeated application of the above operators. Evolution Survival of the fittest You will be evolving generations of Evolution will be guided by fitness… “individuals” “Individuals” are solutions to your You must decide what makes a “fit” problem solution Are you a good solution or a bad solution. 4
Genotype / Phenotype Solution breeding Genotype (DNA) of your solution will be Child solutions are spawned from a common data structure parent solutions List, array, tree, etc. Crossover Mechanism to combine DNA to create DNA Phenotype will be an “individual” of offspring solution derived from the DNA You must define this translation. Solution mutation Solution mutation The DNA from a single individual can The DNA from a single individual can randomly mutate. randomly mutate. Mutation Mutation Must define how given the data structure Must define how given the data structure representing the genotype. representing the genotype. Goal Questions so far To create individuals suitable to a given environment… In other words… Create “fit” individuals Create “good” solutions Create solutions that best solve the problem. 5
Evotionary Computation process Given a problem To use evolutionary algorithms your must: Define your genotype Initialize population Identify your phenotype Select individuals for crossover (based on Define the genotype -> phenotype translation fitness function Crossover Define crossover and mutation operators Define fitness Mutation Determine selection criteria Insert new offspring Set population parameters into population This course will explore each part of the Are stopping criteria satisfied? process in depth! Finish Lectures Course deliverables Lectures will cover: In this course you will: Process of using evolutionary algorithms Choose a hard, multidimensional problem to solve. Process applied to my favorite problem (TSP) Define and code evolutionary methods to solve the problem. Example uses of EAs in practice Experiment with the parameters. Listed as APPLICATION in schedule. Guest lecturers. You will be working in teams of 2. Teaming up Course deliverables Four major deliverables: Find a common problem… Checkpoints / assignments Implementation code Mycourses interest questionaire: Final report problem/area you would like to solve. Presentation Please fill in by Monday. I will list students and topics on Web by And for grad students: next class. Grad report. 6
Checkpoints Report Weekly (approx) assignments, each exploring Final findings of how well your EC on task in using EC on your problem: solved your problem. 1. Framework 2. Problem Statement Summary of findings in checkpoint 3. Genotype / Phenotype 4. Crossover / Mutation 5. Fitness / Selection 6. Population Parameters Code Presentation Implementation of the EC: Oral summary of your final report From scratch Last weeks of class + finals week. Using a toolkit / API (some listed on Web Peer evaluation. page) Use language of your choice Readable code please Grad report Grading About grad work Graduate Undergrad Help you with project / thesis Assignments 25% 30% Code 20% 25% Report can be the background/previous Report 20% 25% work section of your grad report Presentation 15% 20% Choose problem related to your work. Grad Report 20% 7
Ground rules Due dates Use the language / toolkit of your Checkpoints -- periodically (approx choice weekly) Toolkit use is optional Code + Report: Work in teams of 2. Presentations: week 9, 10, finals week All submissions will be done Grad report: finals week electronically using mycourses dropboxes. Questions Next time: EA Frameworks Defining your problem Please think about your problem and fill out mycourses questionairre. 8
Recommend
More recommend
