parent selection strategies for evolutionary algorithms a comparison of parent selection strategies...

16
Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Upload: collin-bridges

Post on 17-Jan-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Parent Selection Strategies for Evolutionary Algorithms

A Comparison of Parent Selection Strategies Modeled After Human Social InteractionBy Michael Ames

Page 2: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Motivation

Because existing research is limited, but shows significant improvement in algorithm robustness, additional research is warranted.

Evolutionary algorithms modeled after higher level human social interaction could produce significant improvement in convergence times and individual fitness.

Page 3: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Background

Assortative mating – mate selection based on the level of genetic similarities between mates.

Disassortative mating – mate selection based on the level of genetic dissimilarities between mates.

Page 4: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Background

Hamming distance - The number of bits which differ between two binary strings.

HD = 3

1 1 0 0 0 1 0 1 1 0

1 1 0 1 0 1 1 0 1 0

Page 5: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Goals

Determine difficulty of implementation.

Determine how well the strategies handle premature convergence.

Measure selective pressure. Rate performance vs. other social

strategies Rate performance vs. existing

strategies

Page 6: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Test Algorithms

N-QueensBinary Knapsack

Page 7: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

8-Queens

Q

Q

Q

Q

Q

Q

Q

Q

Page 8: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Binary Knapsack

0 0 0 1 1 1 0 0 0 0 0 1 1 0 10 1 0 1 0 0 1 0 1 0 0 1 1 0 1

.157.1634.201.084.0657.13.457.0265.167.0954.068.024.24.351.0214

.1307.1953.03.0168.2451.0951.124.1548.028.1641.13.0651.158.2687.1241

1 0 1 1 0 1 1 1 1 1 1 0 0 0 10 1 0 1 0 0 1 0 0 1 1 1 1 0 00 0 1 0 1 0 0 1 0 1 0 1 1 0 11 1 1 0 0 0 0 1 1 0 1 0 1 0 00 1 1 0 0 1 1 0 1 1 0 0 1 0 1

Page 9: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Approach

Implement the evolutionary algorithms

Implement a separate mate selection object.

Page 10: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Results NQueen

-8

-7

-6

-5

-4

-3

-2

-1

0

1000 3000 5000 7000 9000

Tournament

AMEAMarriage

AMEA LowCheat

AMEA HighCheat

DMEAMarriage

DMEA LowCheat

DMEA HighCheat

Page 11: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Results NQueen

Tournament – generations:3925time:115.46 sec

AMEA Marriage – generations: 5316 time:168.43 sec

AMEA Low Cheat – generations: 4876time:142.57 sec

AMEA High Cheat – generations: 4215 time:127.31 sec

Page 12: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Results Knapsack

0

5

10

15

20

25

30

1000 3000 5000

Tournament

AMEAMarriage

AMEA LowCheat

AMEA HighCheat

DMEAMarriage

DMEA LowCheat

DMEA HighCheat

Page 13: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Results Knapsack

Tournament – generations:6975best:27.9844

AMEA Marriage – generations: 6350 best:27.7853

AMEA Low Cheat – generations: 6550best:27.4023

AMEA High Cheat – generations: 6825 best:27.8461

Page 14: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Conclusions

Poor performance Longer time / inferior results

Page 15: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Future Work

Develop other evolutionary mating strategies

Implement one of the algorithms with two separate populations dynamic in size, one designated male the other female, then apply the same or similar strategies

Page 16: Parent Selection Strategies for Evolutionary Algorithms A Comparison of Parent Selection Strategies Modeled After Human Social Interaction By Michael Ames

Any of these?