Optimization model for regional evacuation transportation system using macroscopic productivity function

•We investigate the methodology of using macroscopic fundamental diagram to optimize evacuation network performance under disasters.•We mathematically prove that uniformly metering traffic into network and keeping the network accumulation under a critical value can achieve the best network performance as long as the productivity function is concave.•Two optimal control strategies for short-noticed and no-noticed disasters are developed based on the above findings.

The simulation of mass evacuation traffic processes, while enormously valuable in emergency planning and management, presents a number of challenges to transportation modelers and analysts. One area where evacuation modeling and analysis has lacked is in the ability to determine the specific evacuation travel demand and capacity and conditions under which a road network can most effectively carry the maximum outflow rate for an area under threat of catastrophic disasters. This is a difficult question to answer because evacuations are so complex and can include millions of people, traveling on tens of thousands of miles of roads, lasting several hours or even days in duration. Knowledge of how to reduce the likelihood of over-saturation would be useful, for example, to develop temporally and spatially phased evacuation plans that meter demand into the system for maximum overall benefit. In this paper an optimization model is proposed to maximize evacuation throughput traffic for regional networks. This model aims at optimizing network outflow and trip complete percentage at a macroscopic level by changing the distribution of evacuation traffic in the time horizon. The productivity function, pioneered by Geroliminis and Daganzo (2007, 2008) is used to assess network performance from a macroscopic point of view. Then, an optimization model with the objective of maximizing both total network productivity and outflow rate is proposed. Further, a simulation based study of the New Orleans metropolitan area is used to validate the effectiveness of the optimization model.