A Comparison of Optimization Techniques as Applied to Gain Optimization of a Reactively Loaded Linear Array.
SYRACUSE UNIV N Y DEPT OF ELECTRICAL AND COMPUTER ENGINEERING
Pagination or Media Count:
This report considers the use of reactance loads to maximize the end-fire gain of a linear array of N half-wavelength thin parallel wire dipoles reactively loaded at their centers with only one element externally driven. The use of one feed element eliminates the need for complex transmission line feed networks normally associated with linear arrays. The overall gain realized was between the maximum value obtained by all the elements externally driven with the proper complex currents and the maximum gain obtained using the reactive loads which resonate the real port currents for maximum gain. Four numerical optimization techniques, Univariate, Rosenbrock, DFP Davidon-Fletcher-Powell, and Continued Partan, were tried, and their results as far as correctness and efficiency were tabulated for this class of functions. These four techniques are representative of all currently available numerical optimization methods. For this report, arrays of N 4, 7, and 10 elements ere considered.
- Electrical and Electronic Equipment