Using Evolutionary Algorithms to Design Antennas with Greater Sensitivity to Ultra-High Energy Neutrinos

Genetic Algorithms or GAs are useful tools for efficiently solving problems with high dimensional parameter spaces. GAs are modeled after the biological evolution of species and can be used in a multitude of applications, such as data classification, multivariate regression, and parameter optimization. The GENETIS collaboration uses these algorithms to optimize for a science outcome, specifically, antenna designs for the detection of UHE neutrino-induced radio pulses. We use two different approaches to accomplish this. First, we use the more traditional method of optimizing geometric designs. Using Remcom’s finite-difference time-domain modeling program, XFdtd, we simulate antenna response patterns that are then used in a neutrino in-ice detector simulation to assign a fitness score. Our second approach opts to design the response patterns directly to quantify the potential for improvement of our fitness scores. Here we will show that using these techniques, we were able to create a design with 22% greater sensitivity to in-ice UHE neutrino detection than the VPol antennas currently in use.