TY - GEN
T1 - Localization strategy for island model genetic algorithm to preserve population diversity
AU - Gozali, Alfian Akbar
AU - Fujimura, Shigeru
N1 - Funding Information:
This work was supported by Indonesia Endowment Fund for Education (LPDP), a scholarship from Ministry of Finance, Republic of Indonesia. This work was conducted while at Graduate School of Information, Production, and Systems, Waseda University, Japan.
Publisher Copyright:
© Springer International Publishing AG 2018.
PY - 2018
Y1 - 2018
N2 - Years after being firstly introduced by Fraser and remodeled for modern application by Bremermann, genetic algorithm (GA) has a significant progression to solve many kinds of optimization problems. GA also thrives into many variations of models and approaches. Multi-population or island model GA (IMGA) is one of the commonly used GA models. IMGA is a multi-population GA model objected to getting a better result (aimed to get global optimum) by intrinsically preserve its diversity. Localization strategy of IMGA is a new approach which sees an island as a single living environment for its individuals. An island’s characteristic must be different compared to other islands. Operator parameter configuration or even its core engine (algorithm) represents the nature of an island. These differences will incline into different evolution tracks which can be its speed or pattern. Localization strategy for IMGA uses three kinds of single GA core: standard GA, pseudo GA, and informed GA. Localization strategy implements migration protocol and the bias value to control the movement. The experiment results showed that localization strategy for IMGA succeeds to solve 3-SAT with an excellent performance. This brand new approach is also proven to have a high consistency and durability.
AB - Years after being firstly introduced by Fraser and remodeled for modern application by Bremermann, genetic algorithm (GA) has a significant progression to solve many kinds of optimization problems. GA also thrives into many variations of models and approaches. Multi-population or island model GA (IMGA) is one of the commonly used GA models. IMGA is a multi-population GA model objected to getting a better result (aimed to get global optimum) by intrinsically preserve its diversity. Localization strategy of IMGA is a new approach which sees an island as a single living environment for its individuals. An island’s characteristic must be different compared to other islands. Operator parameter configuration or even its core engine (algorithm) represents the nature of an island. These differences will incline into different evolution tracks which can be its speed or pattern. Localization strategy for IMGA uses three kinds of single GA core: standard GA, pseudo GA, and informed GA. Localization strategy implements migration protocol and the bias value to control the movement. The experiment results showed that localization strategy for IMGA succeeds to solve 3-SAT with an excellent performance. This brand new approach is also proven to have a high consistency and durability.
KW - 3-SAT
KW - Genetic algorithms
KW - Island model genetic algorithm
KW - Localization strategy
UR - http://www.scopus.com/inward/record.url?scp=85020457539&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85020457539&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-60170-0_11
DO - 10.1007/978-3-319-60170-0_11
M3 - Conference contribution
AN - SCOPUS:85020457539
SN - 9783319601694
T3 - Studies in Computational Intelligence
SP - 149
EP - 161
BT - Computer and Information Science
A2 - Lee, Roger
PB - Springer Verlag
T2 - 16th IEEE/ACIS International Conference on Computer and Information Science, ICIS 2017
Y2 - 24 May 2017 through 26 May 2017
ER -