Leveraging social networks for efficient hurricane evacuation

One of the important factors affecting evacuation performance is the departure time choices made by evacuees. Simultaneous departures of evacuees can lead to overloading of road networks causing congestion. We are especially interested in cases when evacuees subject to little or no risk of exposure evacuate along with evacuees subject to higher risk of threat (also known as shadow evacuation). One of the reasons for correlated evacuee departures is higher perceived risk of threat spread through social contacts. In this work, we study an evacuation scenario consisting of a high risk region and a surrounding low risk area. We propose a probabilistic evacuee departure time model incorporating both evacuee individual characteristics and the underlying evacuee social network. We find that the performance of an evacuation process can be improved by forcing a small subset of evacuees (inhibitors) in the low risk area to delay their departure. The performance of an evacuation is measured by both average travel time of the population and total evacuation time of the high risk evacuees. We derive closed form expressions for average travel time for ER random network. A detailed experimental analysis of various inhibitor selection strategies and their effectiveness on different social network topologies and risk distribution is performed. Results indicate that significant improvement in evacuation performance can be achieved in scenarios where evacuee social networks have short average path lengths and topologically influential evacuees do not belong to the high risk regions. Additionally, communities with stronger ties improve evacuation performance.     

Hurricane evacuation demand models with a focus on use for prediction in future events

Although substantial literature exists on understanding hurricane evacuation behavior, few studies have developed models that can be used for predicting evacuation rates in future events. For this paper, we develop new ordered probit models for evacuation using survey data collected in the hurricane-prone state of North Carolina in 2011 and 2012. Since all covariates in the models are available from the census or based on location, the new models can be applied to predict evacuation rates for any future hurricane. The out-of-sample predictive power of the new models are evaluated at the individual household level using cross validation, and the aggregated level using available data from Hurricane Irene (2011), Hurricane Isabel (2003) and Hurricane Floyd (1999). Model results are also compared with an existing participation rate model, and a logistic regression model available from the literature. Results at the individual household level suggests approximately 70% of households’ evacuation behavior will be predicted correctly. Errors are evenly divided between false positives and false negatives, and with accuracy increasing to 100% as the percentage of people who actually evacuate goes to zero or all and decreasing to about 50% when the population is divided and about half of all households actually evacuate. Aggregate results suggest the new models compare favorably to the available ones, with average aggregate evacuation rate errors of five percentage points.     

Joint modeling of evacuation departure and travel times in hurricanes

Hurricanes are costly natural disasters periodically faced by households in coastal and to some extent, inland areas. A detailed understanding of evacuation behavior is fundamental to the development of efficient emergency plans. Once a household decides to evacuate, a key behavioral issue is the time at which individuals depart to reach their destination. An accurate estimation of evacuation departure time is useful to predict evacuation demand over time and develop effective evacuation strategies. In addition, the time it takes for evacuees to reach their preferred destinations is important. A holistic understanding of the factors that affect travel time is useful to emergency officials in controlling road traffic and helps in preventing adverse conditions like traffic jams. Past studies suggest that departure time and travel time can be related. Hence, an important question arises whether there is an interdependence between evacuation departure time and travel time? Does departing close to the landfall increases the possibility of traveling short distances? Are people more likely to depart early when destined to longer distances? In this study, we present a model to jointly estimate departure and travel times during hurricane evacuations. Empirical results underscore the importance of accommodating an inter-relationship among these dimensions of evacuation behavior. This paper also attempts to empirically investigate the influence of social ties of individuals on joint estimation of evacuation departure and travel times. Survey data from Hurricane Sandy is used for computing empirical results. Results indicate significant role of social networks in addition to other key factors on evacuation departure and travel times during hurricanes.

     

The representation and implementation of time-dependent inundation in large-scale microscopic evacuation simulations

Multi-agent simulation has increasingly been used for transportation simulation in recent years. With current techniques, it is possible to simulate systems consisting of several million agents. Such multi-agent simulations have been applied to whole cities and even large regions. In this paper it is demonstrated how to adapt an existing multi-agent transportation simulation framework to large-scale pedestrian evacuation simulation. The underlying flow model simulates the traffic-based on a simple queue model where only free speed, bottleneck capacities, and space constraints are taken into account. The queue simulation, albeit simple, captures the most important aspects of evacuations such as the congestion effects of bottlenecks and the time needed to evacuate the endangered area. In the case of an evacuation simulation the network has time-dependent attributes. For instance, large-scale inundations or conflagrations do not cover all the endangered area at once.These time-dependent attributes are modeled as network change events. Network change events are modifying link parameters at predefined points in time. The simulation framework is demonstrated through a case study for the Indonesian city of Padang, which faces a high risk of being inundated by a tsunami.     

Risk-based spatial zone determination problem for stage-based evacuation operations

This study seeks to determine risk-based evacuation subzones for stage-based evacuation operations in a region threatened/affected by a disaster so that information-based evacuation strategies can be implemented in real-time for the subzone currently with highest evacuation risk to achieve some system-level performance objectives. Labeled the evacuation risk zone (ERZ), this subzone encompasses the spatial locations containing the population with highest evacuation risk which is a measure based on whether the population at a location can be safely evacuated before the disaster impacts it. The ERZ for a stage is calculated based on the evolving disaster characteristics, traffic demand pattern, and network supply conditions over the region in real-time subject to the resource limitations (personnel, equipment, etc.) of the disaster response operators related to implementing the evacuation strategies. Thereby, the estimated time-dependent lead time to disaster impact at a location and the estimated time-dependent clearance time based on evolving traffic conditions are used to compute evacuation risk. This time-unit measure of evacuation risk enables the ERZ concept to be seamlessly applied to different types of disasters, providing a generalized framework for mass evacuation operations in relation to disaster characteristics. Numerical experiments conducted to analyze the performance of the ERZ-based paradigm highlight its benefits in terms of better adapting to the dynamics of disaster impact and ensuring a certain level of operational performance effectiveness benchmarked against the idealized system optimal traffic pattern for the evacuation operation, while efficiently utilizing available disaster response resources.     

Determinants of full and partial household evacuation decision making in hurricane matthew

This paper adds partial household evacuation to the traditional binary evacuate/stay decision. Based on data from a survey of Jacksonville, FL residents after Hurricane Matthew, multinomial (MNL) and random parameter MNL models were developed to determine the influential factors and whether some variables’ effects are more nuanced than prior literature suggests. The random parameter model was preferred to the fixed parameters model. Variables significant in this model included injury concern, certainty about hurricane impact location, age, marital status, family cohesion, and living in mobile or detached homes. Greater injury concern results in lower likelihood of none of the household evacuating and greater likelihood of partial evacuation, but lower likelihood of full household evacuation. Similarly, greater certainty about hurricane impact increased the probability of partial household evacuation but decreased the probability of full evacuation. Respondent age had heterogenous effects; for 85.54% of respondents, additional years of age increased the likelihood of the household staying. Married households had a higher likelihood of staying or evacuating together. Similarly, greater family cohesion was associated with the household remaining together. Living in mobile homes decreased the likelihood that all of the household stays or evacuates and increased the probability of partial household evacuation. Living in a single-family detached home was associated with lower likelihood of all of the household staying or evacuating and a greater likelihood of a partial household evacuation. These findings can inform strategies that influence full or partial household evacuations, material requirements based on these decisions, and ways to reduce family risk.     

Behavior-consistent information-based network traffic control for evacuation operations

The evacuation operations problem aims to avoid or mitigate the potential loss of life in a region threatened or affected by a disaster. It is shaped to a large extent by the evolution of evacuation traffic resulting from the demand–supply interactions of the associated transportation network. Information-based control is a strategic tool for evacuation traffic operations as it can enable greater access to the affected population and more effective response. However, comparatively few studies have focused on the implementation of information-based control in evacuation operations. This study develops a control module for evacuation operations centered on addressing the demand–supply interactions by using behavior-consistent information strategies. These strategies incorporate the likely responses of evacuees to the information provided in the determination of route guidance information. The control module works as an iterative computational process involving an evacuee route choice model and a control model of information strategies to determine the route guidance information to direct evacuation traffic so as to approach a desired network traffic flow pattern. The problem is formulated as a fuzzy logic based optimization framework to explicitly incorporate practical concerns related to information dissemination characteristics and social equity in evacuation operations. Numerical experiments highlight the importance of accounting for the demand–supply interactions, as the use of behavior-consistent information strategies can lead evacuee route choices to approach the operator-desired proportions corresponding to the desired traffic pattern. The results also indicate that while a behavior-consistent information strategy can be effective, gaps with the desired route proportions can exist due to the discrete nature of the linguistic messages and the real-world difficulty in accurately modeling evacuees’ actual route choice behavior.     

A review on travel behaviour modelling in dynamic traffic simulation models for evacuations

Dynamic traffic simulation models are frequently used to support decisions when planning an evacuation. This contribution reviews the different (mathematical) model formulations underlying these traffic simulation models used in evacuation studies and the behavioural assumptions that are made. The appropriateness of these behavioural assumptions is elaborated on in light of the current consensus on evacuation travel behaviour, based on the view from the social sciences as well as empirical studies on evacuation behaviour. The focus lies on how travellers’ decisions are predicted through simulation regarding the choice to evacuate, departure time choice, destination choice, and route choice. For the evacuation participation and departure time choice we argue in favour of the simultaneous approach to dynamic evacuation demand prediction using the repeated binary logit model. For the destination choice we show how further research is needed to generalize the current preliminary findings on the location-type specific destination choice models. For the evacuation route choice we argue in favour of hybrid route choice models that enable both following instructed routes and en-route switches. Within each of these discussions, we point at current limitations and make corresponding suggestions on promising future research directions.     

A threshold model of social contagion process for evacuation decision making

Individual evacuation decisions are often characterized by the influence of one’s social network. In this paper a threshold model of social contagion, originally proposed in the network science literature, is presented to characterize this social influence in the evacuation decision making process. Initiated by a single agent, the condition of a cascade when a portion of the population decides to evacuate has been derived from the model. Simulation models are also developed to investigate the effects of community mixing patterns and the initial seed on cascade propagation and the effect of previous time-steps considered by the agents and the strength of ties on average cascade size. Insights related to social influence include the significant role of mixing patterns among communities in the network and the role of the initial seed on cascade propagation. Specifically, faster propagation of warning is observed in community networks with greater inter-community connections.     

Effects of income inequality on evacuation,reentry and segregation after disasters

Large-scale disasters often trigger mass evacuation due to significant damages to urban systems. Understanding the evacuation and reentry (return) process of affected individuals is crucial for disaster management. Moreover, measuring the heterogeneity in the individuals' post-disaster behavior with respect to their socio-economic characteristics is essential for policy making. Recent studies have used large-scale location datasets collected from mobile devices to analyze post-disaster mobility patterns. Despite the availability of such data and the societal importance of the problem, no studies have focused on how income inequality affects the equity in post-disaster mobility. To overcome these research gaps, we overlay mobility data with income information from census to quantify the effects of income inequality on evacuation and reentry behavior after disasters, and the resulting spatial income segregation. Spatio-temporal analysis using location data of more than 1.7 million mobile phone users from Florida affected by Hurricane Irma reveal significant effects of income inequality on evacuation behavior. Evacuees with higher income were more likely to evacuate from affected areas and reach safer locations with less damage on housing and infrastructure. These differences were common among evacuees from both inside and outside mandatory evacuation zones. As a result of such effects of inequality, significant spatial income segregation was observed in the affected areas. Insights on the effects of income inequality on post-disaster mobility and spatial segregation could contribute to policies that better address social equity in pre-disaster preparation and post-disaster relief.     

Distributed coordinated signal timing optimization in connected transportation networks

This paper presents a Distributed-Coordinated methodology for signal timing optimization in connected urban street networks. The underlying assumption is that all vehicles and intersections are connected and intersections can share information with each other. The novelty of the work arises from reformulating the signal timing optimization problem from a central architecture, where all signal timing parameters are optimized in one mathematical program, to a decentralized approach, where a mathematical program controls the timing of only a single intersection. As a result of this distribution, the complexity of the problem is significantly reduced thus, the proposed approach is real-time and scalable. Furthermore, distributed mathematical programs continuously coordinate with each other to avoid finding locally optimal solutions and to move towards global optimality. We proposed a real-time and scalable solution technique to solve the problem and applied it to several case study networks under various demand patterns. The algorithm controlled queue length and maximized intersection throughput (between 1% and 5% increase compared to the actuated coordinated signals optimized in VISTRO) and reduced travel time (between 17% and 48% decrease compared to actuated coordinated signals) in all cases.     

Mapping evacuation risk on transportation networks using a spatial optimization model

The focus of this paper is on the development of a methodology to identify network and demographic characteristics on real transportation networks which may lead to significant problems in evacuation during some extreme event, like a wildfire or hazardous material spill. We present an optimization model, called the critical cluster model, that can be used to identify small areas or neighborhoods which have high ratios of population to exit capacity. Although this model in its simplest form is a nonlinear, constrained optimization problem, a special integer-linear programming equivalent can be formulated. Special contiguity constraints are needed to keep identified clusters spatially connected. We present details on how this model can be solved optimally as well as discuss computational experience for several example transportation networks. We describe how this model can be integrated within a GIS system to produce maps of evacuation risk or vulnerability. This model is now being utilized in several research projects, in Europe and the US.     

Optimal transit routing with partial online information

This paper studies the routing strategy in a transit network with partial online information at stops. By partial online information, we mean that the arrival time of the incoming transit vehicles is available for a subset of the lines serving a stop. To cope with the partial information assumption, a new routing strategy is proposed and closed form formulae for computing expected waiting times and line boarding probabilities are derived. The proposed strategy unifies existing hyperpath-based transit route choice models that assume either no information or full information. Like many existing models, it ensures optimality when all information is available or the headway is exponentially distributed. The problem of determining the attractive set is discussed for each of the three information cases. In particular, a new heuristic algorithm is developed to generate the attractive set in the partial information case, which will always yield a solution no worse than that obtained without any information. The paper also reveals that, when information is available, an optimal hyperpath may contain cycles. Accordingly, the cause of such cycles is analyzed, and a sufficient condition that excludes cycles from optimal hyperpaths is proposed. Finally, numerical experiments are conducted to illustrate the impact of information availability on expected travel times and transit line load distributions. Among other findings, the results suggest that it is more useful to have information on faster lines than on slower lines.     

Evaluation of different contraflow strategies for hurricane evacuation in Charleston,South Carolina

Abstract

Limited specific evidence is available on the effectiveness of using contraflow as an evacuation traffic management tool. This study was conducted to determine the best combination of strategy options for evacuating Charleston, SC, along route I-26 during the event of a hurricane or other events. PARAMICS microscopic traffic simulator was used to evaluate the impact of each combination of evacuee response timing and traffic control strategy, such as contraflow, with respect to average vehicular travel time and evacuation duration. Analysis revealed the combination of management strategies that created the lowest evacuation durations and travel times for several types of anticipated evacuee responses. Furthermore, a proposed reconfiguration of the I-526/I-26 interchange for contraflow operations produced additional savings in travel times and evacuation durations. These findings support the use of all lanes for contraflow during all evacuations and provide justification to examine a possible reconfiguration of the I-526/I-26 interchange for use during evacuations.     

Reducing bunching with bus-to-bus cooperation

Schedule-based or headway-based control schemes to reduce bus bunching are not resilient because they cannot prevent buses from losing ground to the buses they follow when disruptions increase the gaps separating them beyond a critical value. (Following buses are then overwhelmed with passengers and cannot process their work quick enough to catch up.) This critical gap problem can be avoided, however, if buses at the leading end of such gaps are given information to cooperate with the ones behind by slowing down.This paper builds on this idea. It proposes an adaptive control scheme that adjusts a bus cruising speed in real-time based on both, its front and rear spacings much as if successive bus pairs were connected by springs. The scheme is shown to yield regular headways with faster bus travel than existing control methods. Its simple and decentralized logic automatically compensates for traffic disruptions and inaccurate bus driver actions. Its hardware and data requirements are minimal.     

A dynamic behavioural traffic assignment model with strategic agents

Foresee traffic conditions and demand is a major issue nowadays that is very often approached using simulation tools. The aim of this work is to propose an innovative strategy to tackle such problem, relying on the presentation and analysis of a behavioural dynamic traffic assignment.The proposal relies on the assumption that travellers take routing policies rather than paths, leading us to introduce the possibility for each simulated agent to apply, in real time, a strategy allowing him to possibly re-route his path depending on the perceived local traffic conditions, jam and/or time already spent in his journey.The re-routing process allows the agents to directly react to any change in the road network. For the sake of simplicity, the agents’ strategy is modelled with a simple neural network whose parameters are determined during a preliminary training stage. The inputs of such neural network read the local information about the route network and the output gives the action to undertake: stay on the same path or modify it. As the agents use only local information, the overall network topology does not really matter, thus the strategy is able to cope with large and not previously explored networks.Numerical experiments are performed on various scenarios containing different proportions of trained strategic agents, agents with random strategies and non strategic agents, to test the robustness and adaptability to new environments and varying network conditions. The methodology is also compared against existing approaches and real world data. The outcome of the experiments suggest that this work-in-progress already produces encouraging results in terms of accuracy and computational efficiency. This indicates that the proposed approach has the potential to provide better tools to investigate and forecast drivers’ choice behaviours. Eventually these tools can improve the delivery and efficiency of traffic information to the drivers.     

Location planning for transit-based evacuation under the risk of service disruptions

The effectiveness of transit-based emergency evacuation highly depends on the location of pick-up facilities, resource allocation, and management. These facilities themselves are often subject to service disruptions during or after the emergency. This paper proposes a reliable emergency facility location model that determines both pre-emergency facility location planning and the evacuation operations afterwards, while facilities are subject to the risk of disruptions. We analyze how evacuation resource availability leverages individual evacuees’ response to service disruptions, and show how equilibrium of the evacuee arrival process could be reached at a functioning pick-up facility. Based on this equilibrium, an optimal resource allocation strategy is found to balance the tradeoff between the evacuees’ risks and the evacuation agency’s operation costs. This leads to the development of a compact polynomial-size linear integer programming formulation that minimizes the total expected system cost from both pre-emergency planning (e.g., facility set-up) and the evacuation operations (e.g., fleet management, transportation, and exposure to hazardous surroundings) across an exponential number of possible disruption scenarios. We also show how the model can be flexibly used to plan not only pre-disaster evacuation but also post-disaster rescue actions. Numerical experiments and an empirical case study for three coastal cities in the State of Mississippi (Biloxi, Gulfport, and D’lberville) are conducted to study the performance of the proposed models and to draw managerial insights.     

Macroscopic relations of urban traffic variables: Bifurcations, multivaluedness and instability

Recent experimental work has shown that the average flow and average density within certain urban networks are related by a unique, reproducible curve known as the Macroscopic Fundamental Diagram (MFD). For networks consisting of a single route this MFD can be predicted analytically; but when the networks consist of multiple overlapping routes experience shows that the flows observed in congestion for a given density are less than those one would predict if the routes were homogeneously congested and did not overlap. These types of networks also tend to jam at densities that are only a fraction of their routes’ average jam density.This paper provides an explanation for these phenomena. It shows that, even for perfectly homogeneous networks with spatially uniform travel patterns, symmetric equilibrium patterns with equal flows and densities across all links are unstable if the average network density is sufficiently high. Instead, the stable equilibrium patterns are asymmetric. For this reason the networks jam at lower densities and exhibit lower flows than one would predict if traffic was evenly distributed.Analysis of small idealized networks that can be treated as simple dynamical systems shows that these networks undergo a bifurcation at a network-specific critical density such that for lower densities the MFDs have predictably high flows and are univalued, and for higher densities the order breaks down. Microsimulations show that this bifurcation also manifests itself in large symmetric networks. In this case though, the bifurcation is more pernicious: once the network density exceeds the critical value, the stable state is one of complete gridlock with zero flow. It is therefore important to ensure in real-world applications that a network’s density never be allowed to approach this critical value.Fortunately, analysis shows that the bifurcation’s critical density increases considerably if some of the drivers choose their routes adaptively in response to traffic conditions. So far, for networks with adaptive drivers, bifurcations have only been observed in simulations, but not (yet) in real life. This could be because real drivers are more adaptive than simulated drivers and/or because the observed real networks were not sufficiently congested.     

Fair dynamic resource allocation in transit-based evacuation planning

Resource allocation in transit-based emergency evacuation is studied in this paper. The goal is to find a method for allocation of resources to communities in an evacuation process which is (1) fair, (2) reasonably efficient, and (3) able to dynamically adapt to the changes to the emergency situation. Four variations of the resource allocation problem, namely maximum rate, minimum clearance time, maximum social welfare, and proportional fair resource allocation, are modeled and compared. It is shown that the optimal answer to each problem can be found efficiently. Additionally, a distributed and dynamic algorithm based on the Lagrangian dual approach, called PFD²A, is developed to find the proportional fair allocation of resources and update the evacuation process in real time whenever new information becomes available. Numerical results for a sample scenario are presented.