Search papers, labs, and topics across Lattice.
This paper introduces a Susceptible-Exposed-Infected-Recovered-Vaccinated (SEIRV) model using ordinary differential equations to analyze malware propagation dynamics and evaluate control strategies. The model incorporates relapse and intervention mechanisms, and the authors analyze the stability of the malware-free equilibrium and identify key parameters influencing the propagation threshold using sensitivity analysis. They further propose a hybrid gradient-based global optimization framework using simulated annealing to identify effective control strategies, and validate the model using a real-world Windows malware dataset, demonstrating the impact of intervention timing on averted infections.
Delaying intervention against malware spread exponentially increases the number of infections, highlighting the critical importance of rapid response.
Malware attacks in today's vast digital ecosystem pose a serious threat. Understanding malware propagation dynamics and designing effective control strategies are therefore essential. In this work, we propose a generic SEIRV model formulated using ordinary differential equations to study malware spread. We establish the positivity and boundedness of the system, derive the malware propagation threshold, and analyze the local and global stability of the malware-free equilibrium. The separatrix defining epidemic regions in the control space is identified, and the existence of a forward bifurcation is demonstrated. Using normalized forward sensitivity indices, we determine the parameters most influential to the propagation threshold. We further examine the nonlinear dependence of key epidemic characteristics on the transmission rate, including the maximum number of infected, time to peak infection, and total number of infected. We propose a hybrid gradient-based global optimization framework using simulated annealing approach to identify effective and cost-efficient control strategies. Finally, we calibrate the proposed model using infection data from the"Windows Malware Dataset with PE API Calls"and investigated the effect of intervention onset time on averted cases, revealing an exponential decay relationship between delayed intervention and averted cases.
