Hidden high-risk states identification from routine urban traffic

Area of Science:

• Complex Systems Science
• Network Science
• Urban Dynamics

Background:

• Identifying hidden high-risk states is vital for preventing system breakdown in complex systems.
• High dimensionality and nonlinear interactions in systems like urban traffic make detecting these states from empirical data challenging, as most states remain unobserved.

Purpose of the Study:

• To develop a method for identifying hidden high-risk states in large-scale complex systems, specifically urban traffic.
• To provide early warning knowledge for risk management by locating unobserved critical states.

Main Methods:

• Inferring the underlying interaction network from urban traffic dynamics using a maximum entropy model.
• Constructing the system's energy landscape to visualize and locate potential high-risk states.
• Utilizing the energy landscape to identify states with a high probability of transitioning to hazardous minima.

Main Results:

• Successfully inferred interaction networks and constructed energy landscapes for urban traffic systems.
• Located hidden high-risk states that were not present in the empirical data.
• These identified states serve as effective risk signals, indicating potential system breakdown.

Conclusions:

• The maximum entropy model and energy landscape approach can effectively identify unobserved high-risk states in complex systems.
• This method provides valuable early warning signals for risk management in urban traffic and potentially other large-scale systems.
• Findings offer insights for proactive risk management strategies to mitigate system breakdown and reduce recovery costs.