A MODIFIED DIFFERENTIAL EVOLUTION ALGORITHM FOR SHAPED BEAM LINEAR ARRAY ANTENNA DESIGN

With the advancement of technology, the need of antenna arrays with shaped power patterns increases day by day for the purpose of improvement of communication. In this article, we represent a new method for designing optimized linear array with shaped beam radiation pattern of desired specifications. The main objective is to obtain suitable current excitation amplitude and phase distribution for the linear array elements so that it can produce the desired custom shaped radiation pattern as the user demands. The design procedure utilizes an improved variant of a prominent and efficient metaheuristics of current interest, namely the Differential Evolution (DE). In our modified DE algorithm, denoted as DE_rBM_2SX, new mutation and crossover strategies are employed. These modifications help to overcome some drawbacks of classical DE. Two examples of linear array with shaped radiation pattern design problem are considered to illustrate the effectiveness of our algorithm. Our results are also compared with two state-of-the-art variants of DE and Particle Swarm Optimization (PSO) - namely JADE and CLPSO (Comprehensive Learning Particle Swarm Optimization). The comparison clearly reveals that our optimization algorithm is more efficient than JADE or CLPSO in finding optimum element excitation amplitude and phase distribution for the desired shaped pattern.