Optimization of Lpda Excitations and the PBM Antenna Benchmarks Using Shade and L-Shade Algorithms

Abstract

The SHADE and L-SHADE variants of the Differential Evolution global search and optimization algorithm are used to compute optimized excitations for a Log Periodic Dipole Array antenna and to numerically solve the Pantoja-Bretones-Martin suite of antenna benchmark problems. Comparison to published data shows that SHADE and L-SHADE both are effective and efficient algorithms for solving the array excitation problem and the Pantoja-Bretones-Martin wire antenna benchmarks. L-SHADE clearly is more efficient on the array problem, but overall on the benchmarks the opposite is true, albeit to a lesser degree. The data support the view that neither algorithm is generally better than the other for the type of wire antenna problems considered here. Rather, which algorithm is more efficient is highly dependent on the specific antenna being optimized. In terms of the quality of their solutions, however, both algorithms accurately return the benchmarks' known global optima while both converge on different optimal array excitations that result in very similar objective function fitnesses.