Transcript Introduction to Evolutionary Computation
Introduction to Evolutionary Computation
Temi avanzati di Intelligenza Artificiale - Lecture 1 Prof. Vincenzo Cutello
Department of Mathematics and Computer Science University of Catania
Evolution
What is Evolution ? Introduction to Evolutionary Computation Lecture 1 2
"Disclaimer"
You may whish to treat this as an abstract idea only It does not matter (in the context of Evolutionary Computation) !
Introduction to Evolutionary Computation Lecture 1 3
Darwinian Evolution
Four Postulates 1.
2.
Individuals within species are variable Some of the variations are passed on to offspring 3.
In every generation, more offspring are produced than can survive 4.
The survival and reproduction of individuals are not random: The individuals who survive and go on to reproduce, or who reproduce the most, are those with the most favourable variations. They are naturally selected.
On the Origin of Species by Means of Natural Selection (Darwin 1859) Introduction to Evolutionary Computation Lecture 1 4
Nature of Natural Selection
Based on "Evolutionary Analysis (Freeman & Herron, 2001)"
Natural Evolution acts...
On Individuals, but the Consequences occur in the population On Individuals, not groups On Phenotypes, but evolution consist of changes in the Genotype On exixting traits, but can produce new traits
Evolution
...
Is backward looking Is not perfect Is nonrandom Is not progressive Introduction to Evolutionary Computation Lecture 1 5
Why are we Interested ?
'Results' of Evolution are 'Creative', 'Surprising', 'Unexpected' 'Highly adapted' to 'Environmental Niches' God or Evolution ?
Can a program 'create things like this' ? Introduction to Evolutionary Computation Lecture 1 6
Why are we interested (contd..) ?
Unsupervised !
No 'conscious' design No knowledge involved Instead: Reproductive Fitness But !
Natural Evolution had an extremely long time (3.7 Billion Years!) Natural Evolution acts in parallel Introduction to Evolutionary Computation Lecture 1 7
Evolutionary Algorithms
Algorithms that are inspired by natural evolution Four Main Elements: Group of Individuals -
Population
Source of Variation -
Genetic Operators
Reproductive Fitness -
Fitness
Survival of the Fittest -
Selection
Search Process Trial and Error Recipe for chosing next trial Introduction to Evolutionary Computation Lecture 1 8
EA Examples 1: Optimization
Airfoil Optimization Other Examples Scheduling Function Optimization Lecture 1 9
EA Examples 2: Exploration
Evolutionary Art Other Examples Electronic Hardware Design Robot Control Introduction to Evolutionary Computation Lecture 1 10
Sex !
Skippers mating, from www.chaparraltree.com/ mn/insects.shtml
Introduction to Evolutionary Computation Lecture 1 11
Some Terms from Genetics
DNA
Very large linear self-replicating molecules found in all living cells, the physical carrier of Genetic Information (Deoxyribonucleic Acid)
Chromosome
A single, very long molecule of DNA
Gene
The basic unit of inheritance, (...) a length of DNA which exerts its influence on an organisms form and function by encoding and directing the synthesis of a protein (...)
Allele
One of a number of alternative forms of a gene that can occupy a given genetic locus on a chromosome.
Introduction to Evolutionary Computation Lecture 1 12
Mutation as a Source of Variation
Mitosis: Nuclear division in Cells Mutations: Errors during Mitosis Point Mutations: simple copy errors - create new alleles Duplication: duplicate stretch of DNA - creates extra genetic material others...
Most Mutations are Neutral !
Introduction to Evolutionary Computation Lecture 1 13
Sexual Reproduction
Additional Steps - Meiosis Combination of chromosome sets from both parents Additional Division Introduction to Evolutionary Computation Lecture 1 14
Recombination in Sexual Reproduction
Mixing of genetic material Mixing chromosomes Mixing genes on single chromosomes (crossover) Creates new combination of existing alleles This is why...
...you can inherit your mother's eyes, and your father's nose Sexual Reproduction Can combine beneficial mutations that arise in different individuals Can elimiate disadvantageous mutations quickly Introduction to Evolutionary Computation Lecture 1 15
Other Aspects of Natural Evolution in EC
Punctuated Equilibrium Viruses Co-Evolution Genetic Engineering Non-Mendelian Inheritance Dominant and Recessive Genes Introduction to Evolutionary Computation Lecture 1 16
Course Overview
Part 1: Basics Representations, Selection, Search Operators Part 2: Other Issues Niching, Co-Evolution, Constraint Handling, Multi Objective Problems, ...
Part 3: Theory Background Knowledge, Basic Results Throughout: Tutorials Tutorials, Exercices, Demos Introduction to Evolutionary Computation Lecture 1 17
References and Resouces for this Lecture
Books Hartl, Daniel L.
Essential Genetics
Jones and Bartlett Publishers, 1996. Introductory genetics text (Barnes Library, q QH 430) (Advanced) Freeman, Scott and Herron, Jon. C. Prentice-Hall 2001. Good book on evolution. (Barnes Library, QH366.2) (Advanced)
Evolutionary Analysis
2nd edition, Stearns, Steven C and Hoekstra, Rolf. F.
Evolution. An Introduction
University Press, 2000. (Barnes Library, QH366.2) (Advanced) Oxford Lawrence, Eleanor
Henderson's Dictionary of Biological Terms
Longman Scientific and Technical, 1989. For Definitions 10th edn. Web Resources
Introduction to evolutionary Biology (Basic)
http://www.talkorigins.org/faqs/faq-intro-to-biology.html
An Introduction to Genetic Analysis Online Book (Advanced)
http://www.ncbi.nlm.nih.gov/books/bv.fcgi?call=bv.View..ShowTOC&rid=iga.T
OC Introduction to Evolutionary Computation Lecture 1 18