LCS Based Diversity Maintenance in Adaptive Genetic Algorithms

View/ Open
File version
Accepted Manuscript (AM)
Author(s)
Ohira, R
Islam, Md Saiful
Jo, jun
Stantic, B
Year published
2019
Metadata
Show full item recordAbstract
A genetic algorithm (GA) experiences premature convergence when the diversity is lost in the population. Adaptive GAs aim to maintain diversity in the population by trading off a balance between exploring the problem space and exploiting known solutions. Existing metrics for population diversity measures only examine the similarity between individuals on a genetic level. However, similarities in the order of genes in individuals in ordered problems, such as the travelling salesman problem (TSP) can play an important role in effective diversity measures. By examining the similarities of individuals by the order of their genes, ...
View more >A genetic algorithm (GA) experiences premature convergence when the diversity is lost in the population. Adaptive GAs aim to maintain diversity in the population by trading off a balance between exploring the problem space and exploiting known solutions. Existing metrics for population diversity measures only examine the similarity between individuals on a genetic level. However, similarities in the order of genes in individuals in ordered problems, such as the travelling salesman problem (TSP) can play an important role in effective diversity measures. By examining the similarities of individuals by the order of their genes, this paper proposes longest common subsequence (LCS) based metrics for measuring population diversity and its application in adaptive GAs for solving TSP. Extensive experimental results demonstrate the superiority of our proposal to existing approaches.
View less >
View more >A genetic algorithm (GA) experiences premature convergence when the diversity is lost in the population. Adaptive GAs aim to maintain diversity in the population by trading off a balance between exploring the problem space and exploiting known solutions. Existing metrics for population diversity measures only examine the similarity between individuals on a genetic level. However, similarities in the order of genes in individuals in ordered problems, such as the travelling salesman problem (TSP) can play an important role in effective diversity measures. By examining the similarities of individuals by the order of their genes, this paper proposes longest common subsequence (LCS) based metrics for measuring population diversity and its application in adaptive GAs for solving TSP. Extensive experimental results demonstrate the superiority of our proposal to existing approaches.
View less >
Conference Title
Communications in Computer and Information Science
Volume
996
Copyright Statement
© Springer Nature Singapore Pte Ltd. 2019. This is the author-manuscript version of this paper. Reproduced in accordance with the copyright policy of the publisher.The original publication is available at www.springerlink.com
Subject
Optimisation
Computational complexity and computability
Artificial life and complex adaptive systems