Design of Antennas from Primitive Shapes Using Evolutionary Algorithms

Astrophysics often prioritizes the detection of faint, hard-to-observe signals. One of the highest goals in this field is to maximize the sensitivity of instruments to increase the probability of discovery. Evolutionary algorithms (EAs) use concepts of biological evolution to find solutions to complex optimization problems. In this work, we present an EA framework to evolve antenna designs from primitive geometric shapes, optimized for desired electromagnetic performance. The algorithm constructs antennas using combinations of basic shapes, which are evaluated through antenna simulation software (Remcom XFdtd). Fitness metrics include both matching prescribed radiation patterns and maximizing sensitivity in user-specified directions. We explore evolved designs for single-feed, multifeed, and array configurations, showcasing performance across both single-frequency and broadband regimes. This work builds upon reflector optimization methods developed for deep-space communication systems and radio astronomy experiments. Future work will incorporate electromagnetic modeling from astrophysical simulation software to enable sensitivity-driven antenna design tailored for next-generation space science missions.