This study first introduces an order-free chromosome encoding to enhance
the performance of genetic algorithms by learning the linkage of building
blocks in non-binary encodings. The method introduces a measure called
affinity which is based on the statistical properties of gene valuations
in the population. It uses the affinity values of the local and global
gene pairs to construct a global permutation with tight building block
positioning. Method is tested and experimental results are shown for a
group of deceptive and real life test problems.
Then, study proposes a gene level concurrency model where each gene
position is implemented on a different process. This combines the
advantages of implicit parallelism and a chromosome structure free
approach. It also helps implementation of gene reordering method
introduced and probably other non-linear chromosome encodings.
Identifer | oai:union.ndltd.org:METU/oai:etd.lib.metu.edu.tr:http://etd.lib.metu.edu.tr/upload/12606188/index.pdf |
Date | 01 August 2002 |
Creators | Sehitoglu, Onur Tolga |
Contributors | Ucoluk, Gokturk |
Publisher | METU |
Source Sets | Middle East Technical Univ. |
Language | English |
Detected Language | English |
Type | Ph.D. Thesis |
Format | text/pdf |
Rights | To liberate the content for public access |
Page generated in 0.0018 seconds