A Resilient Agent-Based Re-Organizing Traffic Network for Urban Evacuations

Implementing effective traffic road reversals is a complex problem: it requires clearing roads from traffic before implementing safe road reversal operations and often results in anomalies in the network topology. Road reversals are further complicated when, due to unexpected events (e.g., torrential rains), roads are suddenly closed. Current traffic road reversal approaches are based on the execution of mathematical models which identify upfront, optimal reversal configurations for the entire traffic network. These approaches assume that the traffic network structure is static, and as such do not allow for dynamic road closures. In this paper, we present a resilient agent-based re-organizing traffic model for urban evacuations. Resilience refers to the traffic network’s ability to regain its evacuation function quickly and efficiently after severe perturbations. The proposed model integrates road reversal and zoning strategies. Experimental results show that: (a) the model improves the evacuation effort, and (b) the evacuation function is able to cope quickly and effectively with dynamic road closures.