Robust Optimization With Worst Case Sensitivity Analysis Applied to Array Synthesis and Antenna Designs

Tolerance analysis studies the impact of tolerances and uncertainties on design performance, which represents a key problem in practical implementation of electromagnetic designs. Building on a worst case error tolerance model, this paper proposes particle swarm optimization (PSO) as an approach of the worst case sensitivity analysis (WCSA). Several methods of WCSA have been compared, such as random search, vertex prediction, genetic algorithm, and PSO. PSO with the absorbing boundary condition is proven to be accurate and efficient by several benchmarks. This paper incorporates WCSA with optimization to offer a method which not only finds the optimized design, but also ensures 100% feasibility of the optimized design under the certain tolerance criteria. Linear and planar array synthesis and resonant antenna geometric design demonstrate that within the maximum tolerance given by the algorithm, the optimum solutions are 100% feasible.