An Integrated Optimal Control System for Emergency Evacuation

Thumbnail Image


umi-umd-4551.pdf (1.43 MB)
No. of downloads: 2875

Publication or External Link






How to effectively control evacuation traffic has emerged as one of the critical research issues in transportation community, due to the unusually high demand surge and the often limited network capacity. This dissertation has developed an integrated traffic control system for evacuation operations that may require concurrent implementation of different control options, including traffic routing, contraflow operation, staged evacuation, and intersection signal control.

The system applies a hierarchical control framework to achieve a trade-off between modeling accuracy and operational efficiency for large-scale network applications. The network-level optimization formulations function to assign traffic to different evacuation corridors, select lane reversal configurations for contraflow operations, and identify the evacuation sequence of different demand zones for staged evacuation. With special constraints to approximate flow interactions at intersections, the formulations have introduced two network enhancement approaches with the aim to capture the real-world operational complexities associated with contraflow operations and staged evacuation.

The corridor-level optimization formulations, taking the network-level decisions as input, function to identify the critical control points and generate the optimal signal timings along the major evacuation corridors. The formulations feature the critical intersection concept to reduce the interference of side-street traffic on arterial evacuation flows. This study has also developed an efficient solution method using the Genetic Algorithm based heuristics along with an embedded macroscopic simulator.

This dissertation has also proposed a revised cell transmission model that aims to capture the complex temporal and spatial interactions of evacuation traffic flows for both levels of optimization formulations. This model can significantly reduce the size of the optimization problem, and yet preserve the ability in effectively modeling network traffic dynamics.

Numerical studies were conducted for each individual control component as well as for the entire integrated control system. The results reveal that the staged evacuation and contraflow strategies generated from the proposed formulations can substantially improve the evacuation efficiency and effectively reduce network congestions. Signal control strategies with the critical intersection concept also outperform the state-of-the-practice evacuation signal plans.