1 travel time reliability during evacuation : the impact of 2 heterogeneous driving behavior 3 4 5
Huizhao Tu,Adam J. Pel,Hao Li,Lijun Sun
2011-01-01
Abstract:1 2 Earlier studies have shown that driving behavior differs strongly at emergency conditions 3 and normal traffic conditions. In this paper, we continue on these findings by 4 investigating how these differences in driving behavior have an impact on travel time 5 reliability. In particular, we focus on the effect of (relatively strong) heterogeneity in the 6 driving behavior. To this end, the microscopic simulation framework S-Paramics is 7 adapted accordingly, and applied to the emergency evacuation network of the Dutch city 8 of Almere. This experimental setup allows a structured and in-depth analysis of the 9 relationship between a number of driving behavior parameters and the emergent travel 10 time reliability. The main findings from this study are thus insightful and directly 11 applicable for evacuation planning and management studies. For instance, it is found that 12 although a reduction in drivers’ mean time-headway and minimum gap acceptance 13 typically improves the overall evacuation time, at the same time this yields less reliable 14 travel times. Also, the reliability of travel times decreases over time resulting in (much) 15 less reliable travel times for those travelers who depart later. And finally, in general, 16 heterogeneity in driving behavior strongly reduces travel time reliability. 17 18 Tu, Pel, Li and Sun 3 INTRODUCTION 1 2 Major traffic accidents, extreme weather conditions, and natural and man-made disasters 3 are a few examples of exceptional events that have a relatively small probability of 4 occurring, yet tend to have a very large impact on the functioning of the transportation 5 system. At the same time, there is a societal dependency on the well-functioning of the 6 transportation system, in particular during emergency situations such as an evacuation. 7 This dependency is emphasized by, for instance, the evacuations preceding hurricanes 8 Katrina and Rita in the U.S. in 2005. Such mass evacuations are becoming increasingly 9 more difficult and resource-consuming as the population and urban development of 10 hazard-prone regions grow faster than the road infrastructure capacity [1],[2]. 11 Furthermore, the success of an evacuation strongly depends on many factors, such as 12 amount of warning time, public preparedness and response time, information and 13 instructions dissemination procedure, evacuation shelters and routes, traffic conditions, 14 dynamic traffic management measures, etc. [3],[4]. This complexity in the underlying 15 processes and the multitude of factors influencing these processes can be dealt with via a 16 model-based approach. An evacuation simulation model is then used for the analysis and 17 planning of a large-scale emergency evacuation [5]. 18 Evacuation analyses generally focus on traffic dynamics and the effect of traffic 19 control measures in order to locate possible bottlenecks and predict evacuation times. For 20 an adequate analysis, the travel behavior and driving behavior of evacuees needs to be 21 modeled with sufficient realism. Prior studies have found that this travel behavior and 22 driving behavior of evacuees differs strongly between evacuation conditions and normal 23 traffic conditions (for travel behavior, see [6], for driving behavior, see [7]). This has led 24 to a number of studies on the impact of these differences. For example, at macroscopic 25 level, Pel et al. [8] considers the effects of expected changes in travelers’ departure time 26 and route choice, and road capacities and free speeds. While, at microscopic level, Tu et 27 al. [9],[10] considers the effects of expected changes in mean driving behavior (relating 28 to aspects such as headway, acceleration, reaction time, etc.) and heterogeneity among 29 drivers. 30 Cova and Johnson [11] show that evacuation travel times strongly differ as to the 31 location of a household (i.e., near or far from the exit) as well as the aggregate departure 32 timing and number of available exits. However, a key factor that has not yet received 33 enough attention is the issue of travel time reliability. That is, how do these behavioral 34 changes that are to be expected during an evacuation situation affect the reliability of 35 travel times. One reason why this is important is that, empirical observations suggest that 36 evacuees show a bias towards using familiar routes and motorways, and that this can be 37 ascribed to the perception of these roads being more reliable [12]. For instance, this is 38 supported by the studies by Dow and Cutter [13] and Lindell and Prater [14] reporting 39 high traffic volumes on the interstate motorways in the evacuations preceding 40 respectively hurricane Floyd and hurricane Katrina despite the availability of alternative 41 routes using rural roads. Also, it has been experimentally shown that the appreciation of 42 reliable travel times leads to travelers departing earlier [15], while at the same time such 43 an early peaked travel demand may worsen traffic conditions and yield a longer overall 44 evacuation time [8],[16]. Finally, apart from these behavioral aspects, travel time 45 Tu, Pel, Li and Sun 4 reliability is generally considered a prominent performance indicator for transport 1 systems [15],[17]. 2 This paper investigates how the expected changes in driving behavior during 3 evacuation conditions impacts the travel time reliability. The results, main findings, and 4 discussions provided here are thus valuable for (I) better understanding the relationship 5 between (heterogeneous) driving behavior and travel time reliability, (II) hypothesizing 6 the impacts of the former on the overall evacuation process, and (III) formulating general 7 recommendations for evacuation planning and management studies regarding how to deal 8 with travel time reliability aspects. To this end, first, the next section reviews a number of 9 studies on driving behavior during evacuation and summarizes the expected differences 10 compared to normal driving behavior. The third section then introduces an appropriate 11 measure to quantify travel time reliability. This measure is used in the following analyses 12 of the impact of evacuation-styled driving behavior on the evacuation times and the travel 13 time reliability. The (quantitative) analyses are conducted on the case study describing 14 the evacuation of the Dutch city of Almere. The final section then concludes with a 15 number of important findings and research implications for future evacuation studies. 16 DRIVING BEHAVIOR DURING EVACUATION 17 18 There is general consensus that the behavior of drivers under mentally demanding and 19 emergency conditions differs from that expressed in normal conditions. This is supported 20 by experimental and empirical findings (e.g., [18]–[23]). Nevertheless, the quantitative 21 changes in driving behavior that can be expected during an evacuation are still subject to 22 debate. Hence, in this paper, we conduct a sensitivity analysis on a range of parameter 23 values (around a parameter set calibrated on normal driving behavior) reflecting the range 24 of expected behavioral changes that are reported in the literature. The latter are shown in 25 Table 1, presenting the observed or assumed changes in a number of driving behavior 26 parameters. 27 TABLE 1 Expected changes in driving behavior during exceptional events 28 Parameter H a m d a r a n d M a h m a ss a n i [1 9 ] H o o g en d o o rn [ 2 0 ] K n o o p e t a l. [ 2 1 ] H o o g en d o o rn e t a l.