Adaptive mutation parameter toggling genetic algorithm for phase-only array synthesis

A genetic algorithm that speeds convergence for phased array phase-only synthesis by adaptively toggling between nine mutation parameter pairs is illustrated. This adaptive algorithm outperforms any of the corresponding nine static cases where these same mutation parameters are held constant throughout the optimisation process.     

Optimization of thinned aperiodic linear phased arrays using genetic algorithms to reduce grating lobes during scanning

The scan volume of a thinned periodic linear phased array is proportional to the spacing between array elements. As the spacing between elements increases beyond a half wavelength, the scan range of the array will be significantly reduced due to the appearance of grating lobes. This paper investigates a method of creating thinned aperiodic linear phased arrays through the application of genetic algorithms that will suppress the grating lobes with increased steering angles. In addition, the genetic algorithm will place restrictions on the driving-point impedance of each element so that they are well behaved during scanning. A genetic algorithm approach is also introduced for the purpose of evolving an optimal set of matching networks. Finally, an efficient technique for evaluating the directivity of an aperiodic array of half-wave dipoles is developed for use in conjunction with genetic algorithms.     

Be/spl acute/zier representations for the multiobjective optimization of conformal array amplitude weights

For conformal phased arrays, generally the excitation amplitude of the array elements must be adjusted in order to maintain low sidelobes as the array is scanned. While the desired phase weights for maximum gain are deterministically set by the array geometry and scan angle, the representation of optimum low sidelobe amplitude weights remains an open problem. Following up on prior work using the efficiency-constrained optimization of a modified Bernstein polynomial for low sidelobe conformal array synthesis, a Be/spl acute/zier surface is shown to provide a good representation of the optimized amplitude weights with a reduced number of parameters, while demonstrating /spl epsiv/-constraint multi-objective optimization of conformal aperture efficiency versus sidelobe level. These results are extended to include a Be/spl acute/zier volume representation for the multiobjective optimization of conformal aperture efficiency versus both sidelobe level and scan angle.     

Particle swarm optimization versus genetic algorithms for phased array synthesis

Particle swarm optimization is a recently invented high-performance optimizer that is very easy to understand and implement. It is similar in some ways to genetic algorithms or evolutionary algorithms, but requires less computational bookkeeping and generally only a few lines of code. In this paper, a particle swarm optimizer is implemented and compared to a genetic algorithm for phased array synthesis of a far-field sidelobe notch, using amplitude-only, phase-only, and complex tapering. The results show that some optimization scenarios are better suited to one method versus the other (i.e., particle swarm optimization performs better in some cases while genetic algorithms perform better in others), which implies that the two methods traverse the problem hyperspace differently. The particle swarm optimizer shares the ability of the genetic algorithm to handle arbitrary nonlinear cost functions, but with a much simpler implementation it clearly demonstrates good possibilities for widespread use in electromagnetic optimization.