Incorporating Household Gathering and Mode Decisions in Large‐Scale No‐Notice Evacuation Modeling. (13th February 2013)

• Record Type: Journal Article
• Authors: Liu, Sirui; Murray‐Tuite, Pamela; Schweitzer, Lisa; Waller, S. Travis; Bhat, Chandra
• Is Part Of: Computer-aided civil and infrastructure engineering. Volume 29:Number 2(2014:Feb.)
• Page Start: 107
• Page End: 122
• Publication Date: 2013-02-13
• Subjects: Civil engineering -- Data processing -- Periodicals; Computer-aided engineering -- Periodicals; 624.0285
• DOI: 10.1111/mice.12008
• Languages: English
• ISSNs: 1093-9687

Abstract

<abstract abstract-type="main"> <title>Abstract</title> <p>Household members tend to evacuate as a unit. However, most engineering‐based evacuation models treat evacuees as independent and separate entities, and overlook the interactions among household members during an evacuation (i.e., gathering children/spouses or uniting with other family members at home). The omission of these behaviors leads to imprecise modeling of evacuation situations. Additionally, transportation mode choice in a no‐notice evacuation has been seldom investigated. We present a framework to incorporate both household‐gathering behavior and mode choice in an emergency into an evacuation model to examine the effects of these two issues on evacuation efficiency and network performance. The framework was tested in the Chicago metropolitan region for two hypothetical incidents with evacuation radii 5 and 25 miles. Models that omit gathering behavior yield dangerously optimistic evacuation times and network congestion levels compared to models that include family interactions. These optimistic estimates are significant for a large‐scale evacuation—the reduction in the number of evacuees who can reach safe zones in a certain time threshold is nearly 50% between the gathering and no‐gathering models. Gathering behavior could also cause distinct effects on network performance for inner and outer areas, the break point of which may be where severe bottlenecks are located. In this study, average travel speed increases on the overall network within 15 miles of the incident location (where downtown Chicago is located), but decreases outside the 15‐mile radius. These results would be completely overlooked without incorporating the gathering behavior yet are critical for evacuation management.</p> </abstract>