Comparative study of broadcasting antenna array optimization using evolutionary algorithms

Abstract

Broadcasting antenna array optimized design involves gain maximization, main lobe down-tilting and null filling. In this study some of the most powerful evolutionary optimization algorithms are applied to this challenging problem: Differential Evolution, Particle Swarm, Invasive Weed, Adaptive Invasive Weed, and the Taguchi method. Evolutionary algorithms use a random search approach together with mechanisms inspired by biological evolution in order to iteratively improve the precision of randomly obtained solutions. Evolutionary algorithms are shown to require very substantial computational resources due to their random search nature. However, they are also very robust in finding a quasi-optimum solution by optimizing an appropriate fitness function. It is demonstrated that the algorithm producing the best fitness, and thus the best solution to the antenna problem, is Invasive Weed Optimization (IWO), followed by Particle Swarm Optimization (PSO) and Differential Evolution (DE), in second place and with similar results.