Thinning Approach Based on Sides Lobe Level Reduction in the Linear Array Antenna Using Dynamic Differential Evolution

This novel work has considered the problem of minimizing the side lobe level of linear array antenna patterns using the concept of thinning. In comparison to all element arrays, thinning approach activated only a fraction portion of the total available number of array elements at the suitable positions to achieve the objectives. The thinning problem has been transformed into a constrained optimization problem with the objective of minimizing the total number of activated array elements while simultaneously satisfying the desired side lobe and first null beam width. A new approach based on dynamic value assignment of mutation factor in the differential evolution has been proposed, which carries the random value assignment for each component of the differential vector. The proposed dynamic mutation factor assignment helped to explore and converge on a better solution compared to the standard approach. Fundamentally Differential evolution explores the solution over a continuous domain; hence a transformation is needed to convert the solution into the binary domain for thinning. The different existing forms of transformation, including sigmoid function-based S-shaped and V-shaped functions, have been considered under probabilistic and threshold-based binary conversion processes. It has been observed that the proposed form of differential evolution has explored a better solution by rounding the obtained real value to the nearest integer value and later assigning the binary value according to closeness to their binary value limit. For the linear array detail, experimental analysis has given and observed that the proposed solution has outperformed not only other variants of differential evolution but also performed better than particle swarm optimization results available in past literature.