Title
A genetic algorithm with a variable-length genotype and embryogeny for microstructured optical fibre design
Abstract
Microstructured optical fibres are a relatively recent advance in fibre technology which guide light by using arrays of air holes which run the length of the fibre. The internal microstructure of optical fibres can be altered to reshape and transform light for use in medical applications, sensing, long distance and local area network high bandwidth communications. Recent progress in the production of polymer fibres allows designs with complex microstructures consisting of hundreds of holes to be manufactured. In this paper we present a generative (embryogenic) representation which can produce symmetric fibre designs with a variable number of holes. The resulting genetic algorithm has the ability to search designs of varying complexity over time, allowing less or more complex designs to be evolved as required. Various aspects of this representation are discussed in light of the supporting genetic algorithm such as as recombination of designs and the conversion of the variable length binary genotype to the phenotype (optical fibre structure). We include some single objective design results for a high-bandwidth optical fibre along with manufactured designs.
Year
DOI
Venue
2006
10.1145/1143997.1144278
GECCO
Keywords
Field
DocType
symmetric fibre design,complex microstructures,guide light,high-bandwidth optical fibre,microstructured optical fibre design,microstructured optical fibre,genetic algorithm,optical fibre structure,fibre technology,variable-length genotype,polymer fibre,optical fibre,complex design,constraint satisfaction,local area network,microstructures
Optical fiber,Constraint satisfaction,Mathematical optimization,Microstructure,Fiber,Computer science,Electronic engineering,Local area network,Genetic algorithm,Binary number,High bandwidth
Conference
ISBN
Citations 
PageRank 
1-59593-186-4
1
0.46
References 
Authors
3
4
Name
Order
Citations
PageRank
Steven Manos1263.77
Leon Poladian2162.72
Peter J. Bentley3963113.11
Maryanne Large431.02