Affiliated with the
Communication & Space
Sciences Laboratory

Nature Inspired Antenna Design

Covariance Matrix Adaptation Evolutionary Strategy

Illustration of the selection process used in the CMA-ES

..: References :..

[1]   M. D. Gregory, Z. Bayraktar, and D. H. Werner, "Fast Optimization of Electromagnetic Design Problems Using the Covariance Matrix Adaptation Evolutionary Strategy," IEEE Transactions on Antennas and Propagation, Vol. 59, No. 4, pp. 1275-1285, April 2011.

ABSTRACT: A new method of optimization recently made popular in the evolutionary computation (EC) community is introduced and applied to several electromagnetics design problems. First, a functional overview of the covariance matrix adaptation evolutionary strategy (CMA-ES) is provided. Then, CMA-ES is critiqued alongside a conventional particle swarm optimization (PSO) algorithm via the design of a wideband stacked-patch antenna. Finally, the two algorithms are employed for the design of small to moderate size aperiodic ultrawideband antenna array layouts (up to 100 elements). The results of the two electromagnetics design problems illustrate the ability of CMA-ES to provide a robust, fast and user-friendly alternative to more conventional optimization strategies such as PSO. Moreover, the ultrawideband array designs that were created using CMA-ES are seen to exhibit performances surpassing the best examples that have been reported in recent literature. Link to Article

[2]   M. D. Gregory, S. V. Martin, and D. H. Werner, "Improved Electromagnetics Optimization: The Covariance Matrix Adaptation Evolutionary Strategy," IEEE Antennas and Propagation Magazine, Vol. 57, No. 3, pp. 48-59, July 2015.

ABSTRACT: The covariance matrix adaptation evolutionary strategy (CMA-ES) is explored here as an improved alternative to well-established algorithms used in electromagnetic (EM) optimization. In the past, methods such as the genetic algorithm (GA), particle swarm optimization (PSO), and differential evolution (DE) have commonly been used for EM design. In this article, we examine and compare the performance of CMA-ES, PSO, and DE when applied to test functions and several challenging EM design problems. Of particular interest is demonstrating the ability of the relatively new CMA-ES to more quickly and more reliably find acceptable solutions compared with those of the more classical optimization strategies. In addition, it will be shown that due to its self-adaptive scheme, CMA-ES is a more user-friendly algorithm that requires less knowledge of the problem for preoptimization configuration. Link to Article