Enhancing Resilience of Parallel Intelligent Systems

As globalization continues to intensify, the frequency of natural disasters increases, and system complexity and interdependencies grow, systems need to develop stronger abilities to withstand such uncertainties and recover from disruptions, which is known as resilience. Accordingly, research on system resilience has gained increasing attention as a critical branch of systems engineering. Traditional resilience studies do not fully apply to more complex and intelligent systems, including parallel intelligent systems. Therefore, we propose a new resilience assessment framework suitable for parallel intelligent systems and various complex systems. Furthermore, as technology advances, nodes within systems are gradually acquiring autonomous capabilities. We believe this presents a significant opportunity for systems to address "unknown unknowns." Our research examines the performance of systems with varying degrees of autonomy when facing attacks. Finally, we summarize the five key aspects that need to be considered to enhance the resilience of parallel intelligent systems: robustness, flexibility, visibility, responsiveness, and autonomy.