Planning for Evacuation: Insights from an Efficient Network Design Model

Abstract

In the context of evacuation planning at a strategic level, an efficient model is proposed to augment network capacity under a budget constraint so as to enhance operational performance under a security threat. Contraflow strategies and lane additions are considered for capacity augmentation. The corresponding mathematical programming model is obtained through a case-specific graph theoretic transformation of Daganzo's cell transmission model, exploiting problem characteristics to enhance computational efficiency. This paper focuses on experiments and sensitivity analyses to provide insights: (1) on the computational efficiency of the proposed model; and (2) for strategies for evacuation planning and operations. Sensitivity analyses are performed for a test network in terms of the budget for contraflow operations, the evacuee population size, and the amount and spatial distribution of origin-destination (O-D) demand. The results suggest that there is a threshold budget beyond which benefits, in terms of the network clearance time, are negligible. They also indicate that the network clearance time varies linearly with uniformly distributed population size. Finally, greater spatial uniformity of O-D demand reduces clearance times, suggesting that specifying multiple destinations in the safety zone may lead to more efficient evacuation in many instances. From a response standpoint, the study suggests that operational effectiveness requires predetermined contraflow strategy plans with designated and well-trained personnel, rather than just reacting to a security event in real-time.