Classification and analysis of cyber-attack intensities: a game theoretic approach

Abstract

Computer networks are integral to modern life, yet face escalating security risks, necessitating innovative cybersecurity strategies. This study introduces a novel noncooperative game theory model to enhance cyber-attack prevention. The model, structured as a two-player game with attackers and defenders, aims to identify and mitigate cybersecurity risks through a three-level strategy approach. The model yields a Mixed Strategies Nash Equilibrium using non-zero-sum game theory and linear algebraic techniques. Simulations conducted in Python, using libraries such as numpy and nashpy and data scraped from three websites, evaluated residual energy, defense success rates, and defensive redundancy. The results show that the level-0 defense strategy completed 10,000 cycles with 100% residual energy. The developed model (Level-M) showed slightly lower performance with 80.21% residual energy, however, the level-2 defense strategy, which had the highest defense success rate, only achieved 39.9% residual energy. Additionally, Level-M exhibited superior performance with a defensive redundancy of 22.43%, compared to 67.74% for level-2, indicating that Level-M effectively allocates resources and avoids unnecessary defensive mechanisms when no attack occurs. This implies that a noncooperative, non-zero-sum approach can improve the system's defense against cyber threats and produce an economical defense mechanism. Notably, these simulation metrics reveal the model's superior efficacy and efficiency in preventing dynamic cyber threats. Extensive simulations validate our model's efficacy, affirming its capacity to offer invaluable insights into proactive cyber defense mechanisms. The study's conclusions underscore the prospective utility of the formulated game theoretic model, which is poised to empower defenders with enhanced system protection and exploitation strategies.