A model for cascading failures with the probability of failure described as a logistic function

Area of Science:

• Network Science
• Complex Systems
• Systems Engineering

Background:

• Traditional cascading failure models assume node removal upon overload.
• Network mitigation measures often prevent immediate node removal.
• Existing models do not fully capture real-world network dynamics under stress.

Purpose of the Study:

• To propose a novel cascading failure model incorporating a logistic function for node failure probability.
• To analyze network robustness considering mitigation effects.
• To identify optimal network parameters for enhanced resilience against cascading failures.

Main Methods:

• Developed a new cascading failure model using a logistic function for overloaded node failure probability.
• Conducted numerical simulations on Barabási-Albert (BA) scale-free networks and a real airport network.
• Compared the proposed model against a linear failure probability model.

Main Results:

• The robustness difference between models depends on initial load distribution and load redistribution.
• Identified optimal load distribution parameters for maximum network robustness.
• Demonstrated that increased average degree enhances network robustness.

Conclusions:

• The proposed logistic function model offers a more realistic representation of network cascading failures.
• Optimal load distribution and network topology (higher average degree) are crucial for robust network design.
• Findings can inform the development of effective mitigation strategies and resilient network architectures.