Logo image
Back
Analysis of Evolutionary Algorithm based Optimization for Cyber Threat Modeling
Conference proceeding

Analysis of Evolutionary Algorithm based Optimization for Cyber Threat Modeling

Joseph Grady Wright V, Hakki Erhan Sevil, Guillermo Francia, Tarek Youssef, Tirthankar Ghosh and Gregory Hall
SoutheastCon 2022 Proceedings, pp.751-756
SoutheastCon 2022 (Mobile, AL, USA, 03/26/2022–04/03/2022)
03/26/2022
DOI: https://doi.org/10.1109/SoutheastCon48659.2022.9764131
Web of Science ID: WOS:000821548200121

Metrics

136 Record Views

Abstract

This study utilizes a type of genetic algorithm (GA) called trait-based heterogeneous populations plus (TbHP+) for evaluating cyber threat model equations. TbHP+ uses immunity and instinct to create a type of memory concept that then offers populations with more efficient directives. Additionally, TbHP+ uses the varying number of individuals based on "traits" and can change them during the search process therefore allowing the size of a population to automatically adapt. The population size is based on specific characteristics such as character fitness and credit for immunity. The TbHP+ model in this study was first tested and compared to the classical GA model regarding minimum error performance for well-known mathematical test functions: Rastrigin's 6th test function, De Jong's 1st Test function, and Schwefel's 7th test function. Further, TbHP+ and classical GA were applied to a cyber threat model, namely the internet worm model. In both analysis, the TbHP+ algorithm performs better than classical GA algorithm, when compared to the minimum error performance in test functions and to the objective function of the internet worm model.

Details

Logo image