Abstract
A multigrid optimisation strategy is introduced to design passive metallic reflectors with corrugated shapes. The strategy is based on using genetic algorithms at multiple grids and shaping the metal sheets, starting from coarse details to fine tunings. This corresponds to a systematic expansion of the related optimisation space, which is explored more efficiently in comparison to a brute-force optimisation without using grid. By employing the multilevel fast multipole algorithm to analyse the electromagnetic problems corresponding to optimisation trials, we obtain accurately designed reflectors that provide focussing abilities with very high performances at single and multiple locations. The designed reflectors are also resistant to fabrication errors with less complex corrugations and simplified reflection mechanisms compared to those found by no-grid optimisation trials.
Highlights
A multigrid optimisation strategy is introduced to design passive metallic reflectors with corrugated shapes
The designed reflectors are resistant to fabrication errors with less complex corrugations and simplified reflection mechanisms compared to those found by no-grid optimisation trials
We showed that a genetic algorithms (GAs) implementation combined with an iterative fast solver based on the multilevel fast multipole algorithm (MLFMA)[26,27,28] can provide effective designs, which have various focussing abilities when illuminated externally
Summary
A multigrid optimisation strategy is introduced to design passive metallic reflectors with corrugated shapes. The strategy is based on using genetic algorithms at multiple grids and shaping the metal sheets, starting from coarse details to fine tunings This corresponds to a systematic expansion of the related optimisation space, which is explored more efficiently in comparison to a brute-force optimisation without using grid. Nature-inspired algorithms, such as particle swarm optimisation methods[19,20] and genetic algorithms (GAs)[21,22], provide a great freedom on the fitness functions, including those for multipurpose applications[5,23,24] These heuristic algorithms can be combined externally with electromagnetic solvers, while, as a drawback, they need relatively large numbers of trials for satisfactory optimisation results. Besides the difficulty in the realisation of these structures, they are observed to be sensitive to fabrication errors
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call