Design and Optimization of a Tri-Band Terahertz Antenna Using a Genetic Algorithm Based on Random Forest Model

Amirali, Abdollah; Zareian-Jahromi, Ehsan

doi:10.22034/joeds.2026.569700.1111

Design and Optimization of a Tri-Band Terahertz Antenna Using a Genetic Algorithm Based on Random Forest Model

Document Type : Original Article

Authors

Abdollah Amirali ¹

Ehsan Zareian-Jahromi ²

¹ Department of Electrical engineering, Shiraz University of Technology, Shiraz, Iran

² Shiraz University of Technology

10.22034/joeds.2026.569700.1111

Abstract

Multi-band terahertz antennas are challenging due to complexity and nonlinear trade-offs between design parameters. The novelty of this paper is the presentation of an intelligent optimization framework based on a random forest surrogate model and a genetic algorithm for the design of a compact tri-band terahertz antenna. The antenna structure is created in two steps from a simple rectangular patch by adding corner notches and a U-shaped slot. To overcome the high computational cost of direct simulations, 240 parametric simulations are first performed in CST, and a random forest model with high accuracy (R² > 0.99) is trained. Then, the genetic algorithm uses this fast model (prediction in less than 1 second) to search for the optimal dimensions. The proposed method reduces the computational cost by more than 90% compared to direct optimization. The final optimized antenna has three resonance bands at frequencies of 1.17, 1.46, and 1.76 THz with |S11|<–10 dB, making it suitable for multispectral sensing applications.

Keywords

Antenna

terahertz

genetic algorithm

random forest

surrogate model