Moth Flame Optimization with Random Forest Surrogate Modeling for Side Lobe Suppression in Linear Antenna Arrays

Authors Avishek Das1, Hemant Patidar2, Avishek Chakraborty3, Surajit Mukherjee1, Vikas Maheshwari4
Affiliations

1Department of ECE, HIT, 721657 Haldia, India

2Department of ECE, Oriental University, Indore, India

3Department of EECE, Gitam University, Bengaluru, India

4Department of ECE, Guru Nanak Institutions Technical Campus, Hyderabad, India

Е-mail avishek.uit0408@gmail.com
Issue Volume 18, Year 2026, Number 5
Dates Received 20 June 2026; revised manuscript received 20 October 2026; published online 28 October 2026
Citation Avishek Das, Hemant Patidar, et al., J. Nano- Electron. Phys. 18 No 5, 05026 (2026)
DOI https://doi.org/10.21272/jnep.18(5).05026
PACS Number(s) 84.40.Ba, 84.40.Ua
Keywords Linear antenna array, Side lobe level, Random forest regressor, Moth flame optimization.
Annotation

This paper presents an efficient procedure for the synthesis of current excitation amplitude weights of linear antenna arrays (LAA) aimed at suppressing side lobe levels (SLL) while preserving the main-beam shape. The approach combines the exploratory strength of the moth flame optimization (MFO) algorithm with a surrogate model constructed using a random forest regressor. The surrogate replaces the majority of full electromagnetic evaluations during the optimization cycle, thereby reducing computational burden without compromising accuracy. The methodology is examined through the design of 10-element and 16-element arrays with half-wavelength spacing. Numerical results confirm that the proposed hybrid strategy provides improved side lobe performance and faster convergence compared with conventional uniform excitation and standalone MFO.

List of References