Dynamic differential evolution strategy and applications in electromagnetic inverse scattering problems

Abstract

A novel evolution algorithm, the dynamic differential evolution strategy, is developed to solve optimization problems. It inherits the genetic operators from the differential evolution strategy. However, it differs itself from the differential evolution strategy by updating its population dynamically while the differential evolution strategy updates its population generation by generation. The dynamic updating of population leads to a larger virtual population size and quicker response to change of population status. Two trial functions have been minimized using the dynamic differential evolution strategy. Comparison with the differential evolution strategy has been carried out. It has been observed that the dynamic differential evolution strategy significantly outperforms the differential evolution strategy in efficiency, robustness, and memory requirement. It is then applied to solve a benchmark electromagnetic inverse scattering problem. The outstanding performance of the dynamic differential evolution strategy is further consolidated.