DM-LIMGA

Dual Migration Localized Island Model Genetic Algorithm—a better diversity preserver island model

Alfian Akbar Gozali, Shigeru Fujimura

Research output: Contribution to journalArticle

Abstract

Island Model Genetic Algorithm (IMGA) is a multi-population based GA. IMGA aimed to avoid local optimum by maintaining population (island) diversity using migration. There are several mechanisms of migration and individual selection such as the best (or worst) individual selection, new naturally inspired evolution model, and dynamic migration policy. Migration can delay island (local) convergence and intrinsically preserve diversity. Ironically, migration is also potential to bring overall island (global) convergence, faster. In a certain generation, the migrated individuals among islands will have similar value (genetic drift). So, this work aims to preserve global diversity better by implementing Localized IMGA (LIMGA) and Dual Dynamic Migration Policy (DDMP). LIMGA creates unique evolution trends by using a different kind of GAs for each island. DDMP is a new migration policy which rules the individual migration. DDMP determines the state of an island according to its diversity and attractivity level. By determining its states, DDMP ensures the individual migrating to the correct island dynamically. We call the combination of LIMGA and DDMP as Dual Migration LIMGA (DM-LIMGA). Our experiments show that DM-LIMGA can preserve the diversity better. As its implication, DM-LIMGA can create a more extensive search space and dominates the results among other solvers.

Original languageEnglish
JournalEvolutionary Intelligence
DOIs
Publication statusPublished - 2019 Jan 1

Fingerprint

Preserver
Island Model
Genetic Models
Islands
Migration
Public Policy
Genetic algorithms
Genetic Algorithm
Genetic Drift
Population Dynamics
Attractivity
Local Convergence

Keywords

  • Diversity preservation
  • Island model genetic algorithm
  • Localization strategy
  • Migration policy

ASJC Scopus subject areas

  • Mathematics (miscellaneous)
  • Computer Vision and Pattern Recognition
  • Cognitive Neuroscience
  • Artificial Intelligence

Cite this

@article{f737384a5f8a483d80da0438a47bb925,
title = "DM-LIMGA: Dual Migration Localized Island Model Genetic Algorithm—a better diversity preserver island model",
abstract = "Island Model Genetic Algorithm (IMGA) is a multi-population based GA. IMGA aimed to avoid local optimum by maintaining population (island) diversity using migration. There are several mechanisms of migration and individual selection such as the best (or worst) individual selection, new naturally inspired evolution model, and dynamic migration policy. Migration can delay island (local) convergence and intrinsically preserve diversity. Ironically, migration is also potential to bring overall island (global) convergence, faster. In a certain generation, the migrated individuals among islands will have similar value (genetic drift). So, this work aims to preserve global diversity better by implementing Localized IMGA (LIMGA) and Dual Dynamic Migration Policy (DDMP). LIMGA creates unique evolution trends by using a different kind of GAs for each island. DDMP is a new migration policy which rules the individual migration. DDMP determines the state of an island according to its diversity and attractivity level. By determining its states, DDMP ensures the individual migrating to the correct island dynamically. We call the combination of LIMGA and DDMP as Dual Migration LIMGA (DM-LIMGA). Our experiments show that DM-LIMGA can preserve the diversity better. As its implication, DM-LIMGA can create a more extensive search space and dominates the results among other solvers.",
keywords = "Diversity preservation, Island model genetic algorithm, Localization strategy, Migration policy",
author = "Gozali, {Alfian Akbar} and Shigeru Fujimura",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/s12065-019-00253-2",
language = "English",
journal = "Evolutionary Intelligence",
issn = "1864-5909",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - DM-LIMGA

T2 - Dual Migration Localized Island Model Genetic Algorithm—a better diversity preserver island model

AU - Gozali, Alfian Akbar

AU - Fujimura, Shigeru

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Island Model Genetic Algorithm (IMGA) is a multi-population based GA. IMGA aimed to avoid local optimum by maintaining population (island) diversity using migration. There are several mechanisms of migration and individual selection such as the best (or worst) individual selection, new naturally inspired evolution model, and dynamic migration policy. Migration can delay island (local) convergence and intrinsically preserve diversity. Ironically, migration is also potential to bring overall island (global) convergence, faster. In a certain generation, the migrated individuals among islands will have similar value (genetic drift). So, this work aims to preserve global diversity better by implementing Localized IMGA (LIMGA) and Dual Dynamic Migration Policy (DDMP). LIMGA creates unique evolution trends by using a different kind of GAs for each island. DDMP is a new migration policy which rules the individual migration. DDMP determines the state of an island according to its diversity and attractivity level. By determining its states, DDMP ensures the individual migrating to the correct island dynamically. We call the combination of LIMGA and DDMP as Dual Migration LIMGA (DM-LIMGA). Our experiments show that DM-LIMGA can preserve the diversity better. As its implication, DM-LIMGA can create a more extensive search space and dominates the results among other solvers.

AB - Island Model Genetic Algorithm (IMGA) is a multi-population based GA. IMGA aimed to avoid local optimum by maintaining population (island) diversity using migration. There are several mechanisms of migration and individual selection such as the best (or worst) individual selection, new naturally inspired evolution model, and dynamic migration policy. Migration can delay island (local) convergence and intrinsically preserve diversity. Ironically, migration is also potential to bring overall island (global) convergence, faster. In a certain generation, the migrated individuals among islands will have similar value (genetic drift). So, this work aims to preserve global diversity better by implementing Localized IMGA (LIMGA) and Dual Dynamic Migration Policy (DDMP). LIMGA creates unique evolution trends by using a different kind of GAs for each island. DDMP is a new migration policy which rules the individual migration. DDMP determines the state of an island according to its diversity and attractivity level. By determining its states, DDMP ensures the individual migrating to the correct island dynamically. We call the combination of LIMGA and DDMP as Dual Migration LIMGA (DM-LIMGA). Our experiments show that DM-LIMGA can preserve the diversity better. As its implication, DM-LIMGA can create a more extensive search space and dominates the results among other solvers.

KW - Diversity preservation

KW - Island model genetic algorithm

KW - Localization strategy

KW - Migration policy

UR - http://www.scopus.com/inward/record.url?scp=85067990886&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85067990886&partnerID=8YFLogxK

U2 - 10.1007/s12065-019-00253-2

DO - 10.1007/s12065-019-00253-2

M3 - Article

JO - Evolutionary Intelligence

JF - Evolutionary Intelligence

SN - 1864-5909

ER -