prepared by jeethan & jun 1. overview evolutionary algorithms (ea) ea’s v/s traditional...

25
Prepared by Jeethan & Jun 1

Upload: junior-henderson

Post on 26-Dec-2015

220 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Prepared by Jeethan & Jun

1

Page 2: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Overview Evolutionary Algorithms (EA) EA’s v/s Traditional search Pseudo code Parameters Characteristics of EAs Types of Eas Advantages and disadvantages References

2

Page 3: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Search Problem

Darwinian natural selection

Evolutionary Algorithms are population-based “generate-and-test” search algorithms

3

Page 4: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

4

Page 5: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Evolutionary algorithms operate on a population of potential solutions applying the principle of survival of the fittest to produce better approximations to a solution.

A type of Guided Random Search Used for optimization problems

5

Page 6: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

search is performed in a parallel manner Provides a number of potential solutions to

a given problem. They are generally more straight forward to

apply The final choice is left to the user

6

Page 7: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

7

Page 8: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Parameters of EAs may differ from one type to another. Main parameters:

◦ Population size◦ Maximum number of generations◦ Elitism factor◦ Mutation rate◦ Cross-over rate

8

Page 9: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

There are six main characteristics of EAs◦ Representation◦ Selection◦ Recombination◦ Mutation◦ Fitness Function◦ Survivor Decision

Representation:

◦ How to define an individual◦ The way to store the optimization parameters.◦ Determined according to the problem.◦ Different types:

Binary representation Real-valued representation Lisp-S expression representation

9

Page 10: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Selection

◦ Selection determines, which individuals are chosen for mating (recombination) and how many offspring each selected individual produces.

◦  Parents are selected according to their fitness by means of one of the following algorithms: Roulette wheel selection Truncation selection

Recombination

◦ Determines how to combine the genes of selected parents

◦ Types is determined according to the representation :

Bits of the genes Values of the genes

10

Page 11: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Mutation◦ Change on a single gene of the individual

Fitness Function◦ Gives an intuition about how good the individual is.

Survivor Decision◦ Idea of survival of the best individuals. It is about

Elitism factor.

11

Page 12: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Genetic Algorithms(GA) – binary strings

Genetic Programming(GP) – expression trees

Evolutionary Strategies(ES) – real-valued vectors

Evolutionary Programming(EP) – finite state machines

Page 13: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Evolutionary Algorithms

Genetic Algorithms

Genetic Algorithms

Genetic Programming

Genetic Programming

Evolutionary ProgrammingEvolutionary Programming

Evolutionary Strategies

Evolutionary Strategies

Page 14: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Optimum parameter – Random strategy Classified as global search heuristics Represented by byte arrays Two requirements• Genetic representation• Fitness function

Condition principal

Page 15: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Finding the best path between two points in "Grid World"

Creatures in world:◦ Occupy a single

cell◦ Can move to

neighboring cells

Goal: Travel from the gray cell to the green cell in the shortest number of steps

Finish

Start

Page 16: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Representation: N=00, E=10, S=11,W=01

00 10 10 00 10 11 10 10 00 00 10 00 10 10 00 01 00 10 00 10 00 10 10 00 10 11 10 10 00 00 10 00 10 10 00 01 00 10 00 10

10 10 00 10 11 10 00 00 01 00 00 10 00 01 00 01 00 10 10 10

00 00 10 10 00 10 10 11 10 11 10 00 10 00 00 00 01 00 10 00

p1 =

p2 =

p2 =

10 10 00 10 11 10 00 00 01 00 00 10 00 01 00 01 00 10 10 10

00 00 10 10 00 10 10 11 10 11 10 00 10 00 00 00 01 00 10 00 p2 =

p2 =

00 00 10 10 00 10 10 11 10 11 10 00 00 01 00 01 00 10 10 10 p1+2 =

p1 = 00 11 01 10 10 00 00 01 00 10 00 10 11 10 00 00 10 00 10 10

p1’ = 00 11 00 10 10 00 10 01 00 10 00 10 11 10 00 00 11 00 10 10

Population

Fitness function Mutation

Selection

Cross over

Page 17: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

find the proper program simple problems – High computation power represented by expression trees mainly operate cross-over mutation only can be applied once

Page 18: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

no fixed representation Only use mutation operation child is determined in a way of mutation So, we can conclude that there are three

steps:◦ Initialize population and calculate fitness values ◦ Mutate the parents and generate new population◦ Calculate fitness values of new generation and

continue from the second step

Page 19: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

mutation is very critical main application areas:

◦ Cellular design problems.◦ Constraint optimization◦ Testing students’ code◦ ......

not widely used

Page 20: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

Mainly use the real-vectors as coding representation

Very flexible Representation: represent floating, real-

vector as well Selection: neighborhood method

◦ plus selection (both parent and child)◦ comma selection (only parent)

Fitness function: objective function values.

Page 21: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

recombination & mutation: use additional parameters sigma represent the mutation amount

three recombination functions:◦ Arithmetic mean of the parents◦ Geometric mean of the parents◦ Discrete cross-over method.

There are many application areas of the ES. Some of them:Optimization of Road Networks◦ Local Minority Game◦ Multi-Criterion Optimization◦ .....

21

Page 22: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

22

Advantages of Ea’s

• Large application domain • Complex search problems• Easy to work in parallel• Robustness

Disadvantages of Ea’s• Adjustment of parameters (trial-and-error) No guarantee for finding optimal solutions in a finite amount of time

Page 23: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

https://www.youtube.com/watch?v=ejxfTy4lI6I

Page 24: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

http://en.wikipedia.org/wiki/Evolutionary_algorithm http://www.geatbx.com/docu/algindex-02.html#TopOfPage http://www.faqs.org/faqs/ai-faq/genetic/part2/section-3.html http://en.wikipedia.org/wiki/Genetic_programming http://alphard.ethz.ch/gerber/approx/default.html http://en.wikipedia.org/wiki/Evolutionary_programming http://en.wikipedia.org/wiki/Genetic_algorithm http://homepage.sunrise.ch/homepage/pglaus/gentore.htm http://www.ads.tuwien.ac.at/raidl/tspga/TSPGA.html http://en.wikipedia.org/wiki/Evolution_strategy

24

Page 25: Prepared by Jeethan & Jun 1.  Overview  Evolutionary Algorithms (EA)  EA’s v/s Traditional search  Pseudo code  Parameters  Characteristics of EAs

?

25