Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A Safety Level Evaluation Model based on Network Analysis: Enhancing Accessibility & Evacuation Safety in Ho Chi Minh City’s Alleyways

A Safety Level Evaluation Model based on Network Analysis: Enhancing Accessibility & Evacuation... JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING https://doi.org/10.1080/13467581.2022.2050378 A Safety Level Evaluation Model based on Network Analysis: Enhancing Accessibility & Evacuation Safety in Ho Chi Minh City’s Alleyways a b b b Tran Thi To Uyen M.N , Arai Takatoshi , Honma Kentaro and Imai Kotaro a b Department of Architecture, Graduate School of Engineering, The University of Tokyo, Japan; Department of Human and Social Systems, Institute of Industrial Science, The University of Tokyo, Japan ABSTRACT ARTICLE HISTORY Received 23 April 2021 In this study, an evaluation model is developed to analyze the safety level of a street network in Accepted 21 February 2022 terms of accessibility for emergency services and evacuation risk for residents, especially for cities experiencing rapid urbanization and densification. The evaluation model is created based KEYWORDS on the network geometry and street width using the Network Voronoi algorithm, and four Alleyway neighborhoods; Ho evaluation variables are developed, namely the accessibility risk, unreachability risk, edge Chi Minh city; network responsibility, and flow capacity. Next, the model is applied to an alleyway neighborhood in analysis; urban safety Ho Chi Minh City, characterized by a labyrinthine mesh and tree-shaped network, and narrow street widths. Finally, improvement interventions, such as adding new links and widening alleys, are implemented in three case studies, and the results are compared in terms of cost, social impact, and safety improvement. The results show that the most efficient improvement strategy is to target the weakest point in the network, except for the flow capacity, which, however, can detect intersections at risk on evacuation routes, which cannot be derived from the network topology. The developed evaluation model is not only useful to analyze the current risk level in the network but is also a powerful tool to evaluate future infrastructure improvement projects. 1. Introduction experienced rapid urbanization and densification. Labyrinthine mesh and tree-shaped street networks, 1.1. Background and purpose characterized by dead-end streets and narrow street The juxtaposition of conservation and modernization widths, are representing a great risk for urban safety, in the urban tissue and its impact on local communities more specifically the access for emergency services has presented a great challenge for local governments and the evacuation safety of residents in case of a and city planners. Especially in cities that have natural or human-made disaster. In the context of CONTACT Tran Thi To Uyen M.N touyen.tran4@gmail.com Cw-701, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan © 2022 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group on behalf of the Architectural Institute of Japan, Architectural Institute of Korea and Architectural Society of China. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 TRAN T. ET AL. urban expansion and spatial development in Vietnam’s the network topology and street width, and therefore cities, the World Bank (2011) states that upgrading can easily be applied in developing countries, where existing neighborhoods is one of the most efficient data availability is scarce. The robustness of the math- ways to improve housing for the urban poor and the ematical model and capacity of the model to analyze lower middle class without leading to gentrification. the safety level of street networks with limited data However, improvement projects of the alley expansion represent a powerful tool for the sustainable develop- movement in Ho Chi Minh City, which started in the ment and safety enhancement of fast developing cities Phú Nhuận District in 1999, have not been based on a across regions and cultures. quantitative analysis. To verify the effectiveness of this method, the devel- In this context, the purpose of this research is to oped model will be applied to a real urban area in Ho develop an evaluation model to analyze and enhance Chi Minh City’s alleyway neighborhoods, which repre- the accessibility and evacuation safety of street net- sent the core element of the city’s urban identity, and works in cities that have experienced rapid urbaniza- according to Gibert and Pham (2016), the urban net- tion and densification, characterized by labyrinthine work of alleyways still houses about 85% of the city mesh and tree-shaped street networks, using a quanti- dwellers. Shaped like a labyrinthine network between tative analysis to achieve a result of higher resolution. the linear axes of the existing urban grid and formed This model supports the decision-making process of during the French colonial period, the alleyways city planners and local governments in policymaking emerged during the uncertain times of the 1950s and and city governance, especially for alleyway upgrading 1960s as part of a migration to the city and a sponta- projects. This model considers the social impact on neous densification and urbanization process. local communities, the economic impact depending Similar urban fabrics can be found in various cities on the size of the affected area, and the improvement with different climates, cultures, and regional features, of urban safety in terms of accessibility and evacuation, such as the historic center of Damascus in Syria and the to compare the projected results. The developed old medina in the historic city of Fez in Morocco; the model is expected to assess the current safety level of latter was described by Johansson (2006) as an irregu- a street network and detect vulnerable locations in lar street network characterized by narrow streets and order to decide where to execute improvement cut deep canyons where car accessibility is impossible interventions. except for a few distributor roads. This structure can A large number of researches have been conducted also be found in many historical East Asian cities, one on urban safety; however, no current research focuses example is the neighborhood of Kyojima in the eastern on the specific spatial constraints of labyrinthine mesh inner city of Tokyo. Rapidly urbanized without any and tree-shaped street networks combined with nar- planning after the Great Kanto earthquake in 1923, row street widths, using network analysis. Furthermore, the constant threat of natural disasters and the lack while numerous studies have investigated urban safety of functional efficiency compatible with the motor age in developed countries, this research is based solely on still remain (Kitahara 2001). Showcasing the same Figure 1.1. Site location in Ward 2, District 5 in Ho Chi Minh City. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 3 Figure 1.2. Analytical area showing the urban fabric of the site. labyrinthine mesh and tree-shaped street networks Section 3.2. introduces specific characteristics of and narrow street widths, the developed model can different nodes, such as the position and distribution be applied in these urban areas. of house nodes and the location of accessible points, This research focuses on an area located in Ho Chi and measures the street network considering its topo- Minh City’s Ward 2 of District 5 (see Figure 1.1). The logical structure, using the network Voronoi algorithm. neighborhood, displayed in Figure 1.2, is located This research is based on the concept that the safety between four main roads. The inner shape of the ana- level of a location in the network depends on two lytical area represents a typical urban fabric, which factors: the accessibility for emergency vehicles and emerged in the 1950s and 1960s, when many refugees the risks that can be encountered on the evacuation from rural areas migrated to the city during the war route, representing situations with two opposite direc- leading to housing shortages and the densification of tions of movement, entering or leaving the network in the alleyway system. This movement and spontaneous a case of emergency. urban development led to the formation of Ho Chi In section 4, these two strategies are further classi- Minh City’s alleyway system, which has lasted until fied into sub-categories and four new evaluation vari- today. ables are developed, namely the accessibility risk in section 4.1, the unreachability risk in section 4.2, the edge responsibility in section 4.3 and the flow capa- 1.2. Research process and outline of the paper city in section 4.4. The evaluation variables are derived from a bottom-up approach, observing the The goal of this research is to create an original evacuation process and emergency response in dense model using network analysis by developing new urban areas with narrow street widths and mesh and evaluation variables to estimate the safety level in tree-shaped street networks, and describes potential the network. The outline of this paper is illustrated in emergency scenarios and appropriate emergency Figure 1.3. response. In section 3.1 the structure of the network data is In section 5, the behavior of the developed safety described, and the data processing method, which evaluation variables is observed on typical network adds new set parameters and calculations, are topologies. explained to prepare the model for the analysis. 4 TRAN T. ET AL. Figure 1.3. Structure of this paper. In section 6, the developed model is applied to Ho researches dealing with accessibility and emergency Chi Minh City, first, evaluating the safety level of the evacuation, as well as urban safety on different scales current situation in section 6.1, then safety improve- and the socio-economic context of urban commu- ment strategies, which modify the network structure nities, in this order. by adding new edges or widening edges, are applied Tan, Hu, and Lin (2015) state that emergency in three case studies in section 6.2. Finally, the results response activity relies on transportation networks, of the improvement interventions are evaluated in highlighting the primary role of urban street networks section 6.3. and concluded in section 7. for emergency response. This calls for a deeper look into network theory models, where a large number of researches deal with network topologies, robustness, 2. Literature review and modification strategies to increase their resilience. Access to emergency services including police, fire, Different strategies have been developed to optimize rescue, and medical care is essential to the overall network robustness, such as edge rewiring through health, safety, and general welfare of any population, degree-preserving modifications under a constrained and the roadway infrastructure system can greatly budget (Chan and Akoglu 2016), first principled manip- affect accessibility to these services (Novak D. and ulation algorithms by edge/node removal or addition, Sullivan J. 2014). Hence, urban safety has always been based on the measure of natural connectivity (Chan, an essential focus for urban planners and govern- Akoglu, and Tong 2014), and an algebraic connectivity ments. This section explains the relation of this study optimization via link addition compared in a topologi- to existing network analysis models, previous cal metric-based, Fiedler vector-based and random link JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 5 addition models (Wang H. and Van Mieghem P.t 2008). knowledge of the stationary environment during a However, in the case of street networks, specific char- normal situation and the event knowledge of the pre- acteristics related to the network, such as the spatial dictable spatial change for fire-fighting purposes, constraints of the urban context and the metric dis- which breaks the spatial connectivity between adja- tance, have to be considered to modify the network cent spaces and consequently, some escape routes are structure. In this regard, this study integrates, in addi- blocked. A link-focused measure based on the network tion to the street network topology, specific para- theory of closeness and connectivity has been devel- meters of the urban structure to the model, such as oped by Novak and Sullivan (2014), quantifying a link’s the location of houses, the length of street segments relative importance in its system-wide contribution to and street widths. emergency service accessibility. Chen, Wu and Hsu Furthermore, while a certain strategy might work (2019) modify the network structure by widening nar- for a given type of graph, it may not work for other row alleys in the old town of Taipei, combining the network topologies. Therefore, it is necessary to assess facility location problem and shortest path problem the behavior of the specific network topology of dense and focusing on minimizing the response time of fire - and traditional urban networks, which combines labyr- fighting operations. inthine mesh and tree-shaped networks. In this regard, However, these studies analyzing accessibility Buhl et al. (2006) studied topological patterns in street through network structure modifications are not suita- networks of self-organized settlements and analyzed ble to analyze the network topology of this study, their efficiency and robustness using the Minimal consisting of mesh and tree-shaped networks and Spanning Tree (MST) and Greedy Triangulation (GT). characterized by narrow street widths. Furthermore, They found shared structural properties with tunneling whereas previous studies assume that every location networks, as both types of networks exhibit a similar in the street network is usually accessible and only relationship between cost and efficiency. Buhl et al. faces a disconnection in case of a disaster, this study (2004) compared the development of such networks includes the evaluation of street networks where cer- with ant galleries and explained their structure by the tain locations are not accessible to emergency vehicles strong spatial constraints under which they grow and at any time due to narrow street widths. the emergent organization of network patterns stu- A number of authors have analyzed emergency died on social insects: a concept that aligns with the service accessibility based on the floating catchment formation of Ho Chi Minh City’s alleyways. Focusing on method (FCA), which methodologically enables the the efficiency and robustness in ant networks of gal- integration of multiple sources of information into a leries, they state that the efficiency of self-organized single one. KC, Corcoran, and Chhetri (2020) employ graphs is reached by increasing the meshedness, that the enhanced two-step floating catchment method is by merging trees. However, this needs further inves- (E2SFCA) to investigate the optimality of fire station tigation to be applied in an urban context, especially locations in relation to changing spatial distribution of combined with narrow street widths. Based on this the population. Xia et al. (2019) propose a model of study, the developed model analyzes the increase of spatio-temporal accessibility to emergency medical meschedness in the alleyway network and indicates services (ST-E2SFCA), considering temporal variation appropriate positions to add new links in the street in population distribution in the greater Tokyo area network. through a large volume of GPS data of millions of Spatial accessibility to emergency services is an users and compare the accessibility over space and important indicator for evaluating the effectiveness of time. public health services (Xia et al. 2019) and its enhance- Both studies on emergency service accessibility ment is a key strategy to help improve emergency consider temporal variation in population distribution response, minimize property loss, and reduce injuries on a different scale, long-term population growth and and deaths (KC, Corcoran, and Chhetri 2020). There real-time population location, as well as the location of have been numerous studies investigating accessibility emergency facilities. However, a large set of data is of emergency services based on network theory, con- necessary to apply these models on an urban area, sidering potential modifications in the network struc- which is difficult to procure in fast-developing cities, ture. The risk of disruptions in the network can have a where data availability is scarce, often not collected or great impact on the accessibility of emergency ser- difficult to obtain. Therefore, this study introduces an vices. Ertugay, Argyroudis, and Düzgün (2016) consider easily operated evaluation model, based solely on net- road closure probabilities in the case of an earthquake work data, such as the street network and street width, in their accessibility analysis, due to ground failure, which can easily be applied in almost in every urban damage to bridges and overpasses, and collapses of area without relying on large sets of data. buildings adjacent to road edges. Tan, Hu, and Lin In terms of evacuation safety, Brachman and (2015) developed an agent-based simulation of build- Dragicevic (2014) developed an emergency evacuation ing evacuation, considering evacuee’s spatial model based on network science accounting for 6 TRAN T. ET AL. biological variables (fear, survival instinct) and social level, combining pedestrian evacuation of residents variables (emergency management, disaster response) and the network characteristics and topology, stres- in addition to the physical variables (transportation sing the importance of the conservation of traditional network, location of evacuees). Because of its ability neighborhoods, especially in fast-developing cities, to examine various emergency scenarios, the study where community ties are deeply embedded in the requires a great number of datasets, such as GIS data, daily life and social structures of residents. In the case to determine the number and spatial location of evac- of Ho Chi Minh City, Gibert and Pham (2016) empha- uees, registered vehicle data to evaluate the traffic flow size the importance of the multifunctionality and and the location of hazards, such as oil refineries. adaptability of the alleyway neighborhoods from a Similarly, Chen et al. (2020) investigate the evacuation socio-economic point of view and their ability to wel- vulnerability in urban areas in a fine-grained spatio- come a multiplicity of activities and adapt to all social temporal scale, overlapping mobile phone location structures. datasets with the road network of Shanghai. While Although many studies have evaluated urban safety, these studies provide valuable concepts to assess the there seems to be no method that can be adapted in complexity of emergency evacuation through ample traditional and dense urban areas, especially focusing data availability, they do not propose a solution to on mesh and tree-shaped network topologies, com- enhance the current evacuation safety of the urban monly found in self-organized urban structures. structure. In this regard, this study evaluated the cur- Furthermore, research about Vietnamese cities and rent evacuation safety level and evaluates the effi - their urban structures from a computational approach ciency of network modifications, such as edge is scarce, which could be due to the difficulty of obtain- widening or adding new links, essential for the devel- ing sufficient information. Therefore, this study intro- opment of dense urban areas. duces an easily operated evaluation model, based In this context, Zuo et al. (2021) propose a method solely on network data; this is a necessity as Ho Chi of improving emergency evacuation effectiveness Minh City is experiencing rapid development and under the restriction of limited available land resources modernization. in high-density areas, which is combined with urban renewal planning, aiming at optimizing the layout of 3. Analysis target and data processing shelters and evacuation passageway. While the safety of the evacuation process is improved in a “micro The model is based on the geometrical structure of update” manner, they do not consider the addition of the street network in order to calculate the safety level new links in the street network, which is a highly of a position in the network, more precisely, the acces- efficient improvement method for mesh and tree- sibility of emergency services and the evacuation shaped networks, dealt with in this study. danger. Oki and Osaragi (2014) focus on wide-area evacua- tion difficulty in case of a major earthquake in a 3.1. Weighted alleyway neighborhood network densely built-up wooden residential area in Tokyo combining a multi-agent simulation model with a An alleyway neighborhood located in Ward 2 of property damage model describing building col- District 5 in Ho Chi Minh City is taken as a case study lapse. However, in the context of rapid urbanized area. The site has a total area of 0.127 km and contains areas, building information is difficult to obtain and a population density of 37,486 pers./km . (Ho Chi Minh building materials vary between regions and cul- City Statistics Office 2019) tures. Therefore, this study proposes a model that The input data representing land plots was created can be applied in less data covered areas, where based on the digital format map provided by the “Ho urban renewal and safety improvement is highly Chi Minh City Urban Planning Information” (Ho Chi needed. Minh City Department of Planning and Architecture). Zhang et al. (2021) research about pedestrian eva- The CAD drawing contains land plot borders enabling cuation modeling and simulation in multi-exit scenar- the measurement of street width attributed to the ios using a social force model, while Chan (2019) drawn edges along with their length. A new edge is stresses the importance of a mandatory building drawn when there is a change in direction or at an inspection scheme and sustainable building mainte- intersection as no curves are drawn in the network and nance for safer and healthier cities, highlighting the all edges represent straight lines. The entrances of present situation of building deterioration in Hong each house, hereinafter referred to as house nodes, Kong. In order to ensure the health and safety of are connected to the street network by an edge traced people, a safe evacuation process has to be ensured from the middle of the land plot border perpendicular in different emergency scenarios, reaching from the to the street and has a set length of 0. A node is architectural building scale to large urban areas. This positioned at each end of an edge. The edge attributes study focuses on evacuation safety on a neighborhood are classified into nine categories. The edges of the JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 7 Figure 3.1. Network with street width categories. main street representing the border of the alleyway are classified as narrow edges, and all edges with a neighborhood, house edges, and the remaining edges width above 3.5 m are classified as wide edges, are classified depending on the street width with a set including safety threshold of 3.5 m, as shown in Figure 3.1. those with the main street attribute. The safety level This dataset was extracted and applied in a Python of a house is equal to its adjacent edge. In the next program. The proposed model was built using step, all the narrow edges are removed from the net- NetworkX, an open-source Python package, by struc- work graph. Setting the intersection nodes between turing the extracted data creating an adjacent list with the alleyway network and the main street, hereinafter the neighboring nodes of each node and an attribute defined as evacuation points, as source nodes and all list containing information of the edge length and nodes in the network as target nodes, all nodes that street width category. can be reached belong to the accessible network, as referred hereinafter, and all others belong to the inac- cessible network. In 3.2. Accessible and inaccessible networks Figure 3.2, the blue links in the network represent the alleys where emergency vehicles can reach, and According to previous research by Lin and Chen (2009), the red links represent the narrow alleys that are the necessary space for fire operations can be esti- impassable for emergency vehicles or wide edges mated by adding the width of the fire engine, which that are not connected to the accessible network, is a maximum of 2.6 m, the firefighter operation space and therefore belong to the inaccessible network. The of 0.4 m, and the space for the switch door, which is 0.5 evacuation points are the target locations where resi- m. Therefore, the sum of 3.5 m should be the minimum dents should go in case of an evacuation. The nodes space for firefighting and is set as the safety threshold where the accessible network switches to the inacces- to analyze the accessibility of emergency vehicles in sible network are defined as accessible points, as they the network. First, all edges with a width below 3.5 m 8 TRAN T. ET AL. Figure 3.2. Accessible and inaccessible networks. represent the furthest point an emergency vehicle can adequate emergency response due to accessibility dif- enter inside the network. These nodes are key for ficulties of emergency vehicles delaying the response calculating and analyzing the accessibility and evacua- time of emergency services. tion risk inside the network in the next steps (See (2) The unreachability risk and (3) edge responsibil- Figure 3.3). ity consider the risk residents face on the evacuation The accessible points are the generators used to route, if a fire or accident were to cause a disruption in calculate the accessibility network voronoi diagram, the network. The unwell connected network and dead- assigning each node on the inaccessible network to end streets cannot offer two-way evacuation route the closest accessible point. The evacuation points are possibilities for residents, representing a great danger the generators used to calculate the evacuation net- in case of an emergency. work voronoi diagram, assigning each node in the net- (4) The flow capacity focuses on the evacuation work to the closest evacuation point as shown in process of residents if a building fire or the spread of Figure 3.4. smoke forces them to evacuate to a safe area. The narrow and labyrinthine network of the alleyways cre- ates a dangerous environment during the evacuation 4. Safety level evaluation model process for residents. While the accessibility risk and unreachability risk This section introduces four evaluation variables of the give us the safety information of a specific location in safety evaluation model, namely the (1) accessibility the network (node), identifying houses at risk, the edge risk, (2) unreachability risk, (3) edge responsibility, and responsibility and flow capacity focus on the passing of (4) flow capacity, as shown in Figure 4.1. a location in the network (edge), analyzing the robust- The four evaluation variables have been developed ness of the network. to react to specific risks criteria and emergency situa- Fire policy and mitigation strategies in developing tions, analyzed based on the available network data. countries are constrained by inadequate information, (1) The accessibility risk focuses on the response of which is mainly due to a lack of capacity and resources emergency services like an ambulance or firetruck to a for data collection, analysis, and modeling (Masoumi, location in the network, in the case of a person need- van L.Genderen, and Maleki 2019). In densely built ing medical care or a building fire. Dead-end streets urban areas, building fires represent a high risk as it and narrow street widths increase the risk by impeding JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 9 Figure 3.3. Accessible and evacuation points. Figure 3.4. Accessibility Network Voronoi diagram (left) and Evacuation Network Voronoi diagram (right). can easily spread to adjacent buildings. In addition, the the consequence of a disaster. First, the accessibility of access to rescue services in case of a medical emer- emergency vehicles, such as a firetruck or ambulance, gency can greatly affect health, injuries or deaths. which is fundamental to respond to the disaster. In this The extraction of the evaluation variables has been case, the location of accessible points is essential to developed from a bottom-up approach. Due to the evaluate the difficulties emergency responders might limited data availability, possible indicators have face in the network, leading to the development of the been derived from the network topology and street first evaluation variable, i.e., the accessibility risk. width in relation to probable emergency scenarios. As Second, the evacuation of residents is essential to indicated in Figure 4.2, two outcomes are defined as ensure their safety after a disaster occurs, thus the 10 TRAN T. ET AL. Figure 4.1. Structure of the safety evaluation model. Figure 4.2. Evaluation variables extraction process. importance of the location of evacuation points in the point, which represents the nearest location accessible network. In this context, residents can be exposed to by a vehicle from that respective node. While the ideal two possible dangers. On one side, the risk of a poten- accessibility model includes the optimization of the loca- tial disruption in the network as a consequence of the tion of emergency facilities and the time to reach each disaster, disconnecting houses from an evacuation house on a city-wide scale, this index focuses on the scale point and leading to the inability of residents to evac- of a local neighborhood to provide emergency response uate to a safe area. On this account, two evaluation services on a community level. This first step is crucial to variables are developed: the unreachability risk, evalu- achieve a safe city-wide access in the future. ating the potential disconnection of a house from an Accessibility risk of node i is given by evacuation point, and the edge responsibility, evaluat- shortest distance from i to the nearest ing the criticality of an edge, if that edge is removed in d ¼ accessible nodeði : inaccessible nodeÞ (1) the network. On the other side, the risk caused by 0ði : accessible nodeÞ people flows in the network, which can lead to con- gestion and delay in the evacuation process, is mea- The accessibility risk d is expected to be proportional to sured by the fourth evaluation variable, the flow the required emergency response time from node i to capacity. While the studied network characteristics the nearest accessible point; the higher the d value, the enhance these risks, the developed evaluation vari- higher the risk. In Figure 5.2 the distance to the nearest ables are able to detect the most vulnerable locations accessible point is represented by a colorbar that and support the improvement of the safety level of the depends on the value of d as shown in the legend of street network. the figure. The dark blue color represents nodes when d is equal to 0, which means houses are located on the accessible network. The color gradation from blue to 4.1. Accessibility risk red represents the increase in distance from a node to To get a grasp of the current safety level in the network in an accessible point, and therefore a higher risk as emer- terms of accessibility, all the house nodes were measured gency responders need more time and have more diffi - depending on the distance to the nearest accessible culties to reach that node in an emergency situation. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 11 Figure 5.2(c) illustrates edges with an edge respon- 4.2. Unreachability risk sibility r equal to or greater than 1, which are located Each house in the area represents a possible location of on tree-shaped parts of the network ending in a cul- a fire or accident that can cause a disconnection in the de-sac. This leads to a higher risk if the network were to network. To simulate all possible scenarios, any one of be interrupted in case of an emergency, i.e., there is no the edges is removed one after another in front of two-way evacuation possibility. The color gradation every house, breaking the connection in the network from green to red shows the increasing number of in that location. By calculating the paths from all house houses disconnected from any evacuation point if nodes to any evacuation point, the unreachability risk that edge is removed. The highest risk is represented u represents how many times a house node j is not in dark red and is located at the root of the tree net- connected to any evacuation point through all possi- work. In a tree topology, even a single point of failure ble scenarios. can disrupt the connection from a node to a safe area. Unreachability risk of house node jð¼ 1; 2; . . . ; nÞ is The hierarchy of the tree branches and the total num- given by ber of houses on the tree network reflect the respon- sibility an edge holds in case of disruption of that edge u ¼ δ ; j jl l¼1 in the network. where 0 1 if house node j is connected 4.4. Flow capacity < to any evacuation node when B C 0@ A: The flow capacity is calculated by simulating an eva- δ ¼ edge lð¼ 1; 2; . . . ; mÞ; adjacent (2) jl to any house; is removed: cuation process from each house in the network to an 1ðotherwiseÞ: evacuation point and is evaluated based on two para- meters, namely edge capacity and bottleneck risk. For n : Total number of house nodes this evaluation variable, the input data of the edges in m : Total number of edges adjacent to any house the network was modified by dividing the edge length If u is equal to 0, house node j has at least a two-way into equal segments of 1.0 m. This index measures the evacuation route to any evacuation point, represented density of evacuees at the peak time of edges and in dark blue in Figure 5.2(b). If u is equal to or greater detects the locations of bottleneck risks, delaying the than 1, house node j is located on a tree-shaped net- evacuation process. work, inside the mesh network of alleyways. The gra- dation of colors from blue-green to red shows the 4.4.1. Edge capacity increase in risk of unreachability as shown in the The edge capacity represents the average pedestrian graph legend. space on an edge at peak time during the evacuation process and is measured by dividing the edge area by the maximum number of residents on that edge dur- 4.3. Edge responsibility ing the evacuation process. Evacuees reach the nearest The edge responsibility r of edge k is defined by the evacuation point using the shortest path. number house nodes that are disconnected from any Edge capacity of edge k is given by evacuation point in all scenarios simulating a discon- c ¼ (4) nection in the network by removing edge k. k maxfP g kt The edge responsibility represents the importance of an edge to provide an evacuation route between a A : Area of edge k house and a safe area. P : number of evacuees on edge k in time t kt Edge responsibility of edge k is given by We have three assumptions for the calculation of P : evacuees are distributed equally among all house r ¼ h ; kt k kj j¼1 (1) nodes , they have the same walking speed and the same evacuation starting time. Based on these where assumptions, we can calculate P by the distance kt if house node j is connected between any house node and edge k. Max t{P } repre- < kt 0 : to any evacuation node when h ¼ (3) sents the maximum number of evacuees on edge k kj edge k is removed during the whole evacuation process. If max t{P } is kt 1ðotherwiseÞ: equal to 0, c is not defined. The edge capacity c k k n : Total number of house nodes represents the danger of congestion; which is The average number of residents per household has been taken as 4.7 based on the population density indicated for District 5 in the Ho Chi Minh City Statistical Yearbook 2018 (Ho Chi Minh City Statistics Office 2019) 12 TRAN T. ET AL. especially high if two evacuation routes merge without the bottleneck risk. The highest risk of flow capacity an increase of street width, i.e., increase of available arises when the area per person of the edge capacity is area per evacuee. In Figure 5.2(d), c is illustrated by a low, as shown in dark blue, and the difference in gradation from dark blue to light blue. The darker the density of the bottleneck risk is high, as shown in blue, the less space per evacuee is available, which dark red. represents a higher risk; the lighter blue means the available space per evacuee on that edge is larger, 5. Model behavior and application on typical which is safer. street network topologies The evaluation model is first applied on a typical street 4.4.2. Bottleneck risk network topologies to observe the behavior of the The bottleneck risk of a node is caused by the decrease in developed model and to investigate how the model area per person between its adjacent edges in the direc- recognizes locations at risk in different network topol- tion of the evacuation path, which creates a bottleneck ogies. Three abstracted paradigms of street networks effect, and can cause delays in the evacuation process. were constructed to demonstrate how the developed Bottleneck risk of edge k is given by evaluation model analyzes distinctive patterns. The b ¼ c c (5) k k k street network paradigms represent an abstract form of the main categories of typical street patterns found C : edge capacity of edge k s adjacent edge in in cities across regions and cultures. the direction of the evacuation path For each paradigm, first the virtual street network is If b is negative, edge k has no bottleneck risk; if b is drawn inside a square measuring 300 × 300 m. Then, the k k equal to or superior to 0, the end node of edge k in the perimeter of the network is classified as main street, direction of the evacuation route has a bottleneck risk. which belongs to the accessible network. In the next As seen in Figure 5.2(d), the bottleneck risk is repre- step, the accessible points are first located on the peri- sented by a gradation of red and proportionate size of meter, hereinafter referred to as case (1). Then, the acces- the dot. The darker and bigger the red dot, the higher sible points are located inside the network, expanding Figure 5.1. Typical street network topologies. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 13 Figure 5.2. Application of the evaluation model on a virtual street network. the accessible network inside the network, referred to as Figure 5.2 (a) shows the accessibility risk for all case (2). Finally, the other street width categories are topologies. In the case ofgrid (1), the distance to an distributed randomly inside the network in both cases. accessible point is increasing continuously towards the As seen in Figure 5.1, the first street network para- center of the network due to the equal distance dis- digm is a regular grid network and represents a recti- tribution between nodes on the grid network. linear network with a pattern of streets running at right Regarding grid (2), the distance to an accessible point angles. The second street paradigm is a degree-3 net- reacts to the location of the accessible points inside work, also referred to as T-type network. This topology the network, but d remains less than 75 m. The acces- has a high proportion of degree 3 nodes. The third sibility risk shows a similar result in the degree-3 net- street paradigm is a tree and dendritic network. work, where the center area of the network is the most 14 TRAN T. ET AL. vulnerable in case (1) and reacts to the expansion of grid structure is predominant in the center and the accessible network in case (2). Regarding tree and becomes less dense towards the outskirts of the dendritic networks, the accessibility risk increases with city. Narrow and sinuous alleyways composed of the total length and total number of houses of the tree tree and mesh networks fill up the spaces of var- network and is the highest at the “leaves” of the tree ious scales between the grid structure. network. This behavior can be observed in both cases, but the risk decreases in case (2). Naturally, the acces- 6.1. Analysis of the current situation sibility risk is lower in case (2) of all topologies due to the extension of the accessible network. The four evaluation variables have been applied to the In terms of unreachability risk and edge responsi- network of the analytical area in Ward 2, District 5 of bility, the evaluation model reports that all u and r are Ho Chi Minh City. The analyses of the network give a j k equal to 0 in the entire network for the grid and clear understanding of the current level of safety and degree-3 topologies due to their high connectivity. In helps to identify the most vulnerable areas in the net- the case of tree and dendritic networks, the unreach- work for each criterion. ability risk is the highest at the “leaves” of the tree The analysis of the Accessibility risk shows that networks and increases with the size of the tree net- currently 74.12% of the houses in the network are work (Figure 5.2 (b), while the edge responsibility is the not accessible by an emergency vehicle, among highest at network roots, connecting a tree network to which 38.33% are located further than 50 m from an the accessible network as seen in Figure 5.2(c). As seen accessible point (Figure 6.2). As can be observed in in Figure 5.2 (d), the application of the flow capacity Figure 6.3, houses located on a tree-shaped network evaluation on all network topologies shows that the are affected by the unreachability risk as they are not result does not depend on the network topology, but provided with a two-way evacuation route. Currently, rather on the distribution and location of segments 34.87% of the houses are at risk of being unable to with different street widths in the network. reach an evacuation point if an accident were to inter- rupt passage through the network. The Unreachability risk increases proportionately with the size of the tree 6. Model application to Ho Chi Minh City network and the number of houses located between a This section introduces the application of the model in house and an evacuation point. a real urban area, an alleyway neighborhood in Ho Chi In terms of edge responsibility, the risk occurs in tree- shaped networks providing the only access between Minh City. In section 6.1, the current safety level of the houses and an evacuation point. The risk level rises street network is analyzed, then improvement strate- gies are applied in three case studies in section 6.2. with an increase in the length and hierarchy layers of the tree-shaped network as well as with the number of First, the methodology, which combines edge widen- houses located on that network. Consequently, the ing and adding new edges in the network, is explained in section 6.2.1, then the strategies to choose the edge responsibility varies in accordance with the hier- archy of the branches in the tree network and the location of edges for improvement are detailed in sec- number of attached houses and increases towards the tion 6.2.2. Finally, the improvement intervention results are described in section 6.2.3, and compared root of the tree network. Here, 21.59% of the current network has a tree shape, creating subnetworks with and discussed in section 6.3, as illustrated in Figure 6.1. branches of various sizes as seen in Figure 6.4. Ho Chi Minh City was chosen as the analytical area for this study; it exhibits a very unique urban As for the flow capacity, illustrated in Figure 6.5, the trajectory after experiencing colonization, decades edge capacity is represented in blue and the bottle- neck risk of nodes is represented in red. The bottleneck of war, socialism, and de-urbanization, followed by risk of nodes with a higher difference in density the national reunification of 1976 and Đổi Mới reforms (Thrift and Forbes 1986). Before the two appears predominantly at intersections, while the edge capacity is influenced by the width of the street cities were merged and renamed Ho Chi Minh City and the number of times an edge appears on an after the reunification of the country, the French laid the structure for Saigon’s urban grid, while evacuation route and is scattered in the network. Cholon’s economic and social structure shaped its urban form. By 1955, a migration movement to 6.2. Case studies of improvement the city created serious housing problems and overcrowding and led to the spontaneous urban 6.2.1. Methodology development, starting the formation of Ho Chi Three case studies were conducted using the simulation Minh City’s typical alleyway system, which has model focusing on improving the safety level in the net- lasted until today. After the war, the migration work by combing two strategies. The first strategy movement continued and led to further densifica - involves adding a new edge in the network. This method tion of the alleyway system. Therefore, the primary changes the network geometry and has a high impact on JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 15 Figure 6.1. Structure of section 6.Model application to Ho Chi Minh City. the shortest path routes and two-way evacuation routes In the case studies, both methods were combined, by linking dead-end streets and creating new loops inside aiming to improve the safety as much as possible by the network. The addition of a new edge in the network keeping the area subject to change, thereby keeping can be highly efficient while having a small impact on the the costs and impact on the local community as low local community. The second strategy involves widening as possible. The total length of added edges in the existing alleyways. This method has a direct effect on the network and the total area to expand for edge accessibility of emergency vehicles by expanding the widening was equal in all the case studies. The accessible network further and also increases the area total area for edge widening was set at 250 m , of edges on the evacuation route. and the total added edge length was set at 64 m 16 TRAN T. ET AL. Figure 6.2. Accessibility risk. Figure 6.3. Unreachability risk. with a width of 3.5 m. For each case study, the area 6.2.2. Strategy for safety level improvement of improvement was gradually expanded and was The choice of edges to add or widen was decided referred to as high concentration intervention area based on different strategies. Three strategies were in case study 1, medium concentration intervention applied for widening edges. area in case study 2, and low concentration interven- tion area in case study 3, as seen in Fig.5.5. (a) The first strategy to select edges to widen Subsequently, the concentration of interventions of involves connecting edges on the inaccessible improvement decreased from case study 1 to 3. network with a street width greater than 3.5 m JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 17 Figure 6.4. Edge responsibility. Figure 6.5. Flow capacity. to the accessible network (see Fig. 2.1). This (c) The third strategy involves connecting two approach is represented in red in Figure 6.6. accessible networks to one another (see (b) The second approach involves extending the Figure 3.2). This approach is represented in accessible network further into the network yellow in Fig.6.6 and has been combined (see Fig. 2.2). This approach is represented in with the addition of new edges as demon- orange in Figure 6.6. In case study 1, segment strated in case study 2 with the new edges (3) combines this approach with the addition of [4] and [5] and in case study 3 with the new the new edges [2], [3] and [4]. edges [2] and [3]. 18 TRAN T. ET AL. The locations for adding new edges and creating new links in the network targets tree networks with the highest risks first (See Figure 6.2 to 6.5). After targeting the most unsafe areas in the network, the focus of adding new edges shifts to medium and small size networks. The biggest tree networks are connected to the mesh network with the new edge [5] in case study 1, the new edges [3] and [4] in case study 2, and the new edge [2] for the tree network with the highest risk and the new edges [4] and [5] for the second most vulnerable tree network in case study 3. 6.2.3. Improvement intervention results The safety enhancement of the improvement interven- tions varies for each case and is described as follows: Compared to the current situation, the accessibility risk improves considerably inside the intervention area in case study 1 as the accessible network is expanded by both, widening edges and adding new edges con- densed in a small area as seen in Figure 6.7. While widening edges have no influence, adding new edges has similar impacts on the unreachability risk and edge responsibility. No high-risk areas remain in any of the case studies, and the total length of tree networks decreases continuously compared to the cur- rent network from case study 1 to 3 as seen in Figure 6.8 and Figure 6.9. Widening edges can improve the edge capacity, and new links can alter the evacuation route, but not all interventions have an impact on the flow capacity. Adding a new edge only has an impact on the flow capacity if this new connection in the network modifies the shortest path between any house node and an evacuation point. For instance, the new link [2] in case study 1, seen as having an edge capacity value in blue in Figure 6.10, not only decreases the length of the evacuation route of seven houses but also modifies the bottleneck risk in that area: in this case, at the intersection of segment (3). Widening edges can improve the edge capacity of that edge and improve or remove the bottleneck risk around that edge. The new edge [6] in case study 2 reduces the bottleneck risk in proximity of that edge but increases the bottle- neck risk in other locations nearby. This is affected by the change in evacuation routes of 11 houses passing through the new link, which shortens the distance between these houses and the nearest evacuation point. In case study 3, the widening of edges in seg- ment (1) removes a bottleneck with a high-risk value, but smaller bottleneck risk locations appear on the intersection of segment (2) (see Figure 6.10). 6.3. Discussion After applying the proposed model on three case Figure 6.6. Added and widened edges in case studies 1, 2, studies, an accumulative database of graphs and and 3. charts was created to compare the results of the JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 19 Figure 6.7. Accessibility risk. Figure 6.8. Unreachability risk. Figure 6.9. Edge responsibility. Figure 6.10. Flow capacity. case studies with the current situation. As seen in studies, respectively. In case study 3, the houses Figure 6.11, compared to the current situation, the located further than 80 m from an accessible point houses in the network that are inaccessible for an declined sharply from 15.98% in the current situation emergency vehicle decreased by 15.87%, 19.58% to 1.68% (Figure 6.12). While the number of accessible and, 13.89% in the first, second, and third case houses increased in all case studies, from the 20 TRAN T. ET AL. Figure 6.11. Comparison of accessible and inaccessible houses in all case studies. Figure 6.12. Distance distribution between inaccessible houses and accessible points for all case studies. perspective of accessibility for emergency vehicles, 51.47% compared to the current situation, leaving the best improvement was observed in case study 2, only 11.56% of the total network length tree-shaped. with an augmentation of 52.65% of the accessible Figure 6.16 shows that case study 3 has the lowest houses compared to the current situation. edge responsibility risk for each value of r . The risk of unreachability decreased continuously Regarding flow capacity, no case study showed a from case study 1 to 3, with a reduction in risk of clear improvement in both parameters, edge capacity, 35.43%, 44.37%, and 52.32% in the first, second and and bottleneck risk. In terms of third case studies, respectively, compared to the cur- case study 2, the number of edges with c less than rent situation (see Figure 6.13). Case study 3 has only 0.15 m increased by 48.6% compared to the current 16.82% of the houses with no two-way evacuation, network; however, it also showed the highest improve- compared to 34.87% in the current situation, and as ment with an increase of 24.29% in the safest category seen in Figure 6.14, exhibits the most improvement for with c greater than 0.6 m (Figure 6.17). The total each value of u . number of bottleneck risks decreased by 7.19% in The addition of new links in the network enabled case study 1, 10.07% in case study 2, and 9.35% in the connection of tree subnetworks to the mesh net- case study 3 compared to the current network as work, thereby providing a two-way evacuation possi- seen in Figure 6.18. bility in these areas. All case studies reduced the edge Figure 6.19 shows the social impact of the case studies responsibility in all risk levels. As seen in Figure 6.15, on the local residents. While alley widening impacts case study 3 showed the strongest improvement by households by reducing the area of their land plot, add- reducing the total length of tree subnetworks by ing new edges might force residents to relocate. While JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 21 Figure 6.13. Comparison of the frequency of houses and the unreachability risk for all case studies. Figure 6.14. Frequency of the level of unreachability risk among houses with no 2-way evacuation route in all case studies. Figure 6.15. Comparison of length of the network and risk of edge responsibility in all case studies. 22 TRAN T. ET AL. Figure 6.16. Comparison of the network length and the edge responsibility risk categories on tree-shaped networks. Figure 6.17. Comparison of the number of edges and the edge capacity. Figure 6.18. Frequency of the bottleneck risks. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 23 Figure 6.19. Comparison of the social impact in all case studies. the number of households forced to relocate were similar gradually with an increase in the number of houses in all case studies, the number of households affected by located between a house and an evacuation point on a alley widening was the highest in case study 2 and the tree-shaped network. In terms of edge responsibility, lowest in case study 3. Case study 3 applied on a low the risk is the highest at the roots of tree networks, and concentration area had the lowest social impact, while there is a parallel increase with the total length of the showing the best improvement for the accessibility risk, tree network and the number of houses located on especially in the high-risk categories and had the lowest that tree network. Regarding flow capacity, the edge unreachability risk and edge responsibility among all case capacity depends on the distribution of street widths studies conducted in this paper. across the network, and the occurrence of bottleneck If the weakest point in the network is targeted for alley risks is higher at intersections. improvement and is chosen as an investment project, it As confirmed in the result of the case studies, the provides the most efficient result of network-wide safety effectiveness of the improvement interventions varies improvement for the accessibility risk, unreachability risk, depending on the evaluation variables and the and edge responsibility, but this type of scenario is not method applied: edge widening or adding an edge. valid for the concept of flow capacity, where modifica - Regarding adding new edges in the network, the tions in the network only have a local result. However, the accessibility risk can only be improved if the added flow capacity can detect which intersections in the net- edge is connected to the accessible network, but if work are actual intersections on the evacuation routes the added edge modifies the evacuation route of resi- and therefore have a bottleneck risk, which cannot be dents, it can improve or deteriorate the flow capacity, derived from the network topology. In Figure 6.5, the depending on whether the evacuees using the short- bottleneck risk appears to be especially high at intersec- est path from a house to the nearest evacuation point tions, but this is not the case for intersections located at are led through more or less crowded edges, as it the border of the Voronoi network cells where residents might increase the duration of the evacuation process. evacuate in opposite directions. In terms of unreachability risk and edge responsibility, adding a new edge transforms a tree network into a mesh network and provides a two-way evacuation, 7. Conclusion which increases the safety the most. An original methodology for evaluating accessibility As for widening edges in the network, with the and evacuation safety using network analysis was cre- applied strategy, the accessibility risk improves when ated and applied in Ho Chi Minh City’s alleyways. the accessible network is extended, but it has no effect The evaluation of the current street network and the on the unreachability risk and edge responsibility. modifications in the case studies disclosed various While improvement interventions have an impact on results and features of the evaluation variables and safety improvement over a wider area in the network improvement methods. Showing the highest risk at for other evaluation variables, in terms of flow capacity, “leaves” of inaccessible tree-shaped networks, the widening edges only has a local effect in the network. methods applied to improve the accessibility risk The prospective intervention projects were com- involve a combination of reducing the maximum dis- pared in terms of safety improvement, economic tance to an accessible point and extending the acces- impact (defined by the area to modify), and social sible network. The Unreachability risk worsens impact (defined by the number of households forced 24 TRAN T. ET AL. to relocate and households having to give away a part Disclosure statement of their land plot). This evaluation model is not only No potential conflict of interest was reported by the author(s). useful to analyze the current risk level in the network but is also a powerful tool for urban planners and governments to manage and estimate the efficiency Funding of future infrastructure improvement projects. This work was supported by JSPS KAKENHI Grant Number JP From an academic point of view, this research is 20H02327. based on existing network analysis models, such as the Network Voronoi algorithm, which already ORCID proved their efficiency in previous researches but evolves by introducing a novel method adapted to Tran Thi To Uyen M.N http://orcid.org/0000-0002-9218- a specific network topology and considering the urban context and narrow street widths. In prac- tice, the structure and safety of self-organized set- References tlements seems complex, and dangerous locations cannot be pointed out intuitively or are often Brachman, M. L., and Dragicevic S. 2014. “A Spatially Explicit overlooked in traditional urban planning Network Science Model for Emergency Evacuations in an Urban Context.” Computers, Environment and Urban Systems approaches. In this regard, the developed model 44: 15–26. doi:10.1016/j.compenvurbsys.2013.10.009. is able to identify vulnerable locations based on Buhl, J., J. Gautrais, N. Reeves, R. V. Solé, S. Valverde, P. Kuntz, network data solely and evaluate future improve- and G. Theraulaz. 2006. “Topological Patterns in Street ments, which is very helpful for urban manage- Networks of Self-organized Urban Settlements.” The ment and sustainable development, especially in European Physical Journal B - Condensed Matter and fast developing cities. Complex Systems 49 (4): 513–522. doi:10.1140/epjb/ e2006-00085-1. This research, however, is subject to several limita- Buhl J., Gautrais J., Solé R.V., Kuntz P., Valverde S., tions that could be addressed in future research. As Deneubourg J.L. & Theraulaz G.2004. “Efficiency and mentioned in section 3.1, in terms of accessibility, one Robustness in Ant Networks of Galleries.” The European limitation of this study is the boundary of the area, Physical Journal B 42 (1): 123–129. doi:10.1140/epjb/ meaning that activities or the location of emergency e2004-00364-9. Chan, D. W. M. 2019. “Sustainable Building Maintenance facilities outside the analytical area, are not considered for Safer and Healthier Cities: Effective Strategies for in the model. Moreover, this analysis is based on build- Implementing the Mandatory Building Inspection ing fires and medical emergencies and does not include Scheme (MBIS) in Hong Kong.” - Journal of Building other natural or human-made disasters that can occur in Engineering 24: 100737. doi:10.1016/j.jobe.2019.100737. dense urban areas. Another limitation is, that this Chan, H., and Akoglu, L. 2016. “Optimizing Network research is based on network analysis, more specifically Robustness by Edge Rewiring: A General Framework.” Data Mining and Knowledge Discovery 30 (5): 1395–1425. the network Voronoi algorithm, which is based on the doi:10.1007/s10618-015-0447-5. assumption of the shortest path rule. Moreover, this Chen, J., T. Pei, M. Li, C. Song, T. Ma, F. Lu, and S. Shaw. 2020. assumption is based on a number of other assumptions, “An Enhanced Model for Evacuation Vulnerability for example, that all residents are familiar with the Assessment in Urban Areas.” Computers, Environment and whole network structure and positions of locations. Urban Systems 84: 101540. doi:10.1016/j. compenvurbsys.2020.101540. Furthermore, it simplifies the complexity of human deci- Chen, P., H. Wu, and Y. Hsu. 2019. “Widening Narrow Alleys to sions made in case of an emergency, which can be Enhance Response Efficiency for Fire Emergency from the unrealistic in certain situations. Finally, the safety thresh- Perspective of Urban Roadway Network Analysis .” Journal old for the access of emergency vehicles, set at 3.5 m, is of the Eastern Asia Society for Transportation Studies 13 only based on the street width and does not consider 2598–2613 . the geometry of the space such as corners and turning Chen, Jiayan, Jia Yu, Jiahong Wen, Chuanrong Zhang, Zhan’e Yin, Jianping Wu, and Shenjun Yao. 2019. “Pre-evacuation space for emergency vehicles. Time Estimation Based Emergency Evacuation Simulation The next step would be to expand the model to a in Urban Residential Communities.” International Journal city-wide scale and include the locations of emergency of Environmental Research and Public Health 16 (23): 4599. facilities to evaluate the accessibility and response doi:10.3390/ijerph16234599. time of rescuers more precisely. Furthermore, in Chan H., Akoglu L. ,and Tong H.2014. “Make It or Break It: terms of edge responsibility, the extent of overlapping Manipulating Robustness in Large Networks.” In Proceedings of the 2014 SIAM International Conference on parts of two-way evacuation routes should be verified. Data Mining, SDM’14 April 24-26, 2014 Philadelphia, Finally, the current model analyzes the four evaluation Pennsylvania, USA, pp. 325–333 variables independently, which should be combined in Ertugay, K., Argyroudis, S., and Düzgün, H. Ş. 2016. future research to grasp the full potential of improve- “Accessibility Modeling in Earthquake Case considering ment interventions. Road Closure Probabilities: A Case Study of Health and JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 25 Shelter Service Accessibility in Thessaloniki, Greece.” Oki, Takuya, and Toshihiro Osaragi. 2014. “A Method for International Journal of Disaster Risk Reduction 17: 49–66. Evaluation of Wide-area Evacuation Difficulty in case of A doi:10.1016/j.ijdrr.2016.03.005. Major Earthquake.” 12th International Conference on Gibert Marie and Pham Thai Son (2016) “Understanding Design and Decision Support Systems in Architecture and the Vietnamese Urban Fabric from the Inside”. Urban Planning, DDSS 2014 25-27 August 2014 Eindhoven, Producing and living the city in Vietnam, The Focus, Netherlands The Newsletter, No. 73. Tan, L., M. Hu, and Lin, H. 2015. “Agent-based Simulation of Ho Chi Minh City Department of Planning and Architecture, Building Evacuation: Combining Human Behavior with Ho Chi Minh City Urban Planning Information. Application Predictable Spatial Accessibility in a Fire Emergency.” for webpage and mobile devices. Accessed 4 11 2019 Information Sciences 295: 53–66. doi:10.1016/j. https://thongtinquyhoach.hochiminhcity.gov.vn/ ins.2014.09.029. Ho Chi Minh City Statistics Office. 2019. “Ho Chi Minh City Thrift, N., and D. Forbes. 1986. The Price of War: Urbanization Statistical Yearbook 2018”. Population, Labor and employ- in Vietnam, 1954-1985. London: Allen and Unwin. ment in 2018, table 12, Area, population and population Wang, Huijuan, and Piet Van Mieghem. 2008. “Algebraic density in 2018 by district. Connectivity Optimization via Link Addition.” In Johansson, E. 2006. “Influence of Urban Geometry on Proceedings of the 3rd International Conference on Bio- Outdoor Thermal Comfort in A Hot Dry Climate: A Study Inspired Models of Network, Information and Computing in Fez, Morocco.” Building and Environment 41 (10, Systems 25-28 November, 2008 Hyogo, Japan (22). ICST October): 1326–1338. doi:10.1016/j.buildenv.2005.05.022. (Institute for Computer Sciences, Social-Informatics and KC, K., J. Corcoran, and Chhetri, P. 2020. “Measuring the Telecommunications Engineering). Spatial Accessibility to Fire Stations Using Enhanced World Bank (2011) “Vietnam Urbanization Review, Technical Floating Catchment Method.” - Socio-Economic Planning Assistance Report.” In Chapter 3: Urban Expansion and Sciences 69: 100673. doi:10.1016/j.seps.2018.11.010. Spatial Development in Vietnam’s Cities, 3.1 Neighborhood Kitahara, Tomoji. 2001. “Conserving the Walkable Environment Urban Upgrading in HCMC, p.124. in the Neighborhood: A Case Study of the Improvement Xia, T., Song, X., Zhang, H., Song, X., Kanasugi, H., & Effect in Kyojima, Tokyo.” Australia: Walking the 21st Century, Shibasaki, R. 2019. “Measuring Spatio-temporal International Conference, 2001, Perth, Western Australia, 2 Accessibility to Emergency Medical Services through Big Lin, Chen. 2009. “A Study of Fire-fighting Strategy on Narrow GPS Data.” Health & Place 56: 53–62. doi:10.1016/j. Alleys Structures.” Journal of Disaster Mitigation and Rescue healthplace.2019.01.012. (Taiwan) 10 (1): 107–125. Zhang, D., Huang, G., Ji, C., Liu, H., & Tang, Y. 2021. Masoumi, Z., van L.Genderen, J., & Maleki, J. 2019. “Fire Risk “Pedestrian Evacuation Modeling and Simulation in Assessment in Dense Urban Areas Using Information Multi-exit Scenarios.” Physica A 582: 126272. doi:10.1016/ Fusion Techniques.” ISPRS International Journal of Geo- j.physa.2021.126272. Information 8 (12): 579. doi:10.3390/ijgi8120579. Zuo, J., J. Shi, C. Li, T. Mu, Y. Zeng and Dong, J. 2021. Novak D.C., and J.L. Sullivan. 2014. “A Link-focused “Simulation and Optimization of Pedestrian Evacuation in Methodology for Evaluating Accessibility to Emergency High-density Urban Areas for Effectiveness Improvement.” Services.” Decision Support Systems 57 (January): 309–319. - Environmental Impact Assessment Review 87: 106521. doi:10.1016/j.dss.2013.09.015. doi:10.1016/j.eiar.2020.106521. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Asian Architecture and Building Engineering Taylor & Francis

A Safety Level Evaluation Model based on Network Analysis: Enhancing Accessibility & Evacuation Safety in Ho Chi Minh City’s Alleyways

Download PDF
25 pages

Loading next page...
 
/lp/taylor-francis/a-safety-level-evaluation-model-based-on-network-analysis-enhancing-U0Jwo0d92G

References (28)

Publisher
Taylor & Francis
Copyright
© 2022 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group on behalf of the Architectural Institute of Japan, Architectural Institute of Korea and Architectural Society of China.
ISSN
1347-2852
eISSN
1346-7581
DOI
10.1080/13467581.2022.2050378
Publisher site
See Article on Publisher Site

Abstract

JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING https://doi.org/10.1080/13467581.2022.2050378 A Safety Level Evaluation Model based on Network Analysis: Enhancing Accessibility & Evacuation Safety in Ho Chi Minh City’s Alleyways a b b b Tran Thi To Uyen M.N , Arai Takatoshi , Honma Kentaro and Imai Kotaro a b Department of Architecture, Graduate School of Engineering, The University of Tokyo, Japan; Department of Human and Social Systems, Institute of Industrial Science, The University of Tokyo, Japan ABSTRACT ARTICLE HISTORY Received 23 April 2021 In this study, an evaluation model is developed to analyze the safety level of a street network in Accepted 21 February 2022 terms of accessibility for emergency services and evacuation risk for residents, especially for cities experiencing rapid urbanization and densification. The evaluation model is created based KEYWORDS on the network geometry and street width using the Network Voronoi algorithm, and four Alleyway neighborhoods; Ho evaluation variables are developed, namely the accessibility risk, unreachability risk, edge Chi Minh city; network responsibility, and flow capacity. Next, the model is applied to an alleyway neighborhood in analysis; urban safety Ho Chi Minh City, characterized by a labyrinthine mesh and tree-shaped network, and narrow street widths. Finally, improvement interventions, such as adding new links and widening alleys, are implemented in three case studies, and the results are compared in terms of cost, social impact, and safety improvement. The results show that the most efficient improvement strategy is to target the weakest point in the network, except for the flow capacity, which, however, can detect intersections at risk on evacuation routes, which cannot be derived from the network topology. The developed evaluation model is not only useful to analyze the current risk level in the network but is also a powerful tool to evaluate future infrastructure improvement projects. 1. Introduction experienced rapid urbanization and densification. Labyrinthine mesh and tree-shaped street networks, 1.1. Background and purpose characterized by dead-end streets and narrow street The juxtaposition of conservation and modernization widths, are representing a great risk for urban safety, in the urban tissue and its impact on local communities more specifically the access for emergency services has presented a great challenge for local governments and the evacuation safety of residents in case of a and city planners. Especially in cities that have natural or human-made disaster. In the context of CONTACT Tran Thi To Uyen M.N touyen.tran4@gmail.com Cw-701, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan © 2022 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group on behalf of the Architectural Institute of Japan, Architectural Institute of Korea and Architectural Society of China. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 TRAN T. ET AL. urban expansion and spatial development in Vietnam’s the network topology and street width, and therefore cities, the World Bank (2011) states that upgrading can easily be applied in developing countries, where existing neighborhoods is one of the most efficient data availability is scarce. The robustness of the math- ways to improve housing for the urban poor and the ematical model and capacity of the model to analyze lower middle class without leading to gentrification. the safety level of street networks with limited data However, improvement projects of the alley expansion represent a powerful tool for the sustainable develop- movement in Ho Chi Minh City, which started in the ment and safety enhancement of fast developing cities Phú Nhuận District in 1999, have not been based on a across regions and cultures. quantitative analysis. To verify the effectiveness of this method, the devel- In this context, the purpose of this research is to oped model will be applied to a real urban area in Ho develop an evaluation model to analyze and enhance Chi Minh City’s alleyway neighborhoods, which repre- the accessibility and evacuation safety of street net- sent the core element of the city’s urban identity, and works in cities that have experienced rapid urbaniza- according to Gibert and Pham (2016), the urban net- tion and densification, characterized by labyrinthine work of alleyways still houses about 85% of the city mesh and tree-shaped street networks, using a quanti- dwellers. Shaped like a labyrinthine network between tative analysis to achieve a result of higher resolution. the linear axes of the existing urban grid and formed This model supports the decision-making process of during the French colonial period, the alleyways city planners and local governments in policymaking emerged during the uncertain times of the 1950s and and city governance, especially for alleyway upgrading 1960s as part of a migration to the city and a sponta- projects. This model considers the social impact on neous densification and urbanization process. local communities, the economic impact depending Similar urban fabrics can be found in various cities on the size of the affected area, and the improvement with different climates, cultures, and regional features, of urban safety in terms of accessibility and evacuation, such as the historic center of Damascus in Syria and the to compare the projected results. The developed old medina in the historic city of Fez in Morocco; the model is expected to assess the current safety level of latter was described by Johansson (2006) as an irregu- a street network and detect vulnerable locations in lar street network characterized by narrow streets and order to decide where to execute improvement cut deep canyons where car accessibility is impossible interventions. except for a few distributor roads. This structure can A large number of researches have been conducted also be found in many historical East Asian cities, one on urban safety; however, no current research focuses example is the neighborhood of Kyojima in the eastern on the specific spatial constraints of labyrinthine mesh inner city of Tokyo. Rapidly urbanized without any and tree-shaped street networks combined with nar- planning after the Great Kanto earthquake in 1923, row street widths, using network analysis. Furthermore, the constant threat of natural disasters and the lack while numerous studies have investigated urban safety of functional efficiency compatible with the motor age in developed countries, this research is based solely on still remain (Kitahara 2001). Showcasing the same Figure 1.1. Site location in Ward 2, District 5 in Ho Chi Minh City. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 3 Figure 1.2. Analytical area showing the urban fabric of the site. labyrinthine mesh and tree-shaped street networks Section 3.2. introduces specific characteristics of and narrow street widths, the developed model can different nodes, such as the position and distribution be applied in these urban areas. of house nodes and the location of accessible points, This research focuses on an area located in Ho Chi and measures the street network considering its topo- Minh City’s Ward 2 of District 5 (see Figure 1.1). The logical structure, using the network Voronoi algorithm. neighborhood, displayed in Figure 1.2, is located This research is based on the concept that the safety between four main roads. The inner shape of the ana- level of a location in the network depends on two lytical area represents a typical urban fabric, which factors: the accessibility for emergency vehicles and emerged in the 1950s and 1960s, when many refugees the risks that can be encountered on the evacuation from rural areas migrated to the city during the war route, representing situations with two opposite direc- leading to housing shortages and the densification of tions of movement, entering or leaving the network in the alleyway system. This movement and spontaneous a case of emergency. urban development led to the formation of Ho Chi In section 4, these two strategies are further classi- Minh City’s alleyway system, which has lasted until fied into sub-categories and four new evaluation vari- today. ables are developed, namely the accessibility risk in section 4.1, the unreachability risk in section 4.2, the edge responsibility in section 4.3 and the flow capa- 1.2. Research process and outline of the paper city in section 4.4. The evaluation variables are derived from a bottom-up approach, observing the The goal of this research is to create an original evacuation process and emergency response in dense model using network analysis by developing new urban areas with narrow street widths and mesh and evaluation variables to estimate the safety level in tree-shaped street networks, and describes potential the network. The outline of this paper is illustrated in emergency scenarios and appropriate emergency Figure 1.3. response. In section 3.1 the structure of the network data is In section 5, the behavior of the developed safety described, and the data processing method, which evaluation variables is observed on typical network adds new set parameters and calculations, are topologies. explained to prepare the model for the analysis. 4 TRAN T. ET AL. Figure 1.3. Structure of this paper. In section 6, the developed model is applied to Ho researches dealing with accessibility and emergency Chi Minh City, first, evaluating the safety level of the evacuation, as well as urban safety on different scales current situation in section 6.1, then safety improve- and the socio-economic context of urban commu- ment strategies, which modify the network structure nities, in this order. by adding new edges or widening edges, are applied Tan, Hu, and Lin (2015) state that emergency in three case studies in section 6.2. Finally, the results response activity relies on transportation networks, of the improvement interventions are evaluated in highlighting the primary role of urban street networks section 6.3. and concluded in section 7. for emergency response. This calls for a deeper look into network theory models, where a large number of researches deal with network topologies, robustness, 2. Literature review and modification strategies to increase their resilience. Access to emergency services including police, fire, Different strategies have been developed to optimize rescue, and medical care is essential to the overall network robustness, such as edge rewiring through health, safety, and general welfare of any population, degree-preserving modifications under a constrained and the roadway infrastructure system can greatly budget (Chan and Akoglu 2016), first principled manip- affect accessibility to these services (Novak D. and ulation algorithms by edge/node removal or addition, Sullivan J. 2014). Hence, urban safety has always been based on the measure of natural connectivity (Chan, an essential focus for urban planners and govern- Akoglu, and Tong 2014), and an algebraic connectivity ments. This section explains the relation of this study optimization via link addition compared in a topologi- to existing network analysis models, previous cal metric-based, Fiedler vector-based and random link JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 5 addition models (Wang H. and Van Mieghem P.t 2008). knowledge of the stationary environment during a However, in the case of street networks, specific char- normal situation and the event knowledge of the pre- acteristics related to the network, such as the spatial dictable spatial change for fire-fighting purposes, constraints of the urban context and the metric dis- which breaks the spatial connectivity between adja- tance, have to be considered to modify the network cent spaces and consequently, some escape routes are structure. In this regard, this study integrates, in addi- blocked. A link-focused measure based on the network tion to the street network topology, specific para- theory of closeness and connectivity has been devel- meters of the urban structure to the model, such as oped by Novak and Sullivan (2014), quantifying a link’s the location of houses, the length of street segments relative importance in its system-wide contribution to and street widths. emergency service accessibility. Chen, Wu and Hsu Furthermore, while a certain strategy might work (2019) modify the network structure by widening nar- for a given type of graph, it may not work for other row alleys in the old town of Taipei, combining the network topologies. Therefore, it is necessary to assess facility location problem and shortest path problem the behavior of the specific network topology of dense and focusing on minimizing the response time of fire - and traditional urban networks, which combines labyr- fighting operations. inthine mesh and tree-shaped networks. In this regard, However, these studies analyzing accessibility Buhl et al. (2006) studied topological patterns in street through network structure modifications are not suita- networks of self-organized settlements and analyzed ble to analyze the network topology of this study, their efficiency and robustness using the Minimal consisting of mesh and tree-shaped networks and Spanning Tree (MST) and Greedy Triangulation (GT). characterized by narrow street widths. Furthermore, They found shared structural properties with tunneling whereas previous studies assume that every location networks, as both types of networks exhibit a similar in the street network is usually accessible and only relationship between cost and efficiency. Buhl et al. faces a disconnection in case of a disaster, this study (2004) compared the development of such networks includes the evaluation of street networks where cer- with ant galleries and explained their structure by the tain locations are not accessible to emergency vehicles strong spatial constraints under which they grow and at any time due to narrow street widths. the emergent organization of network patterns stu- A number of authors have analyzed emergency died on social insects: a concept that aligns with the service accessibility based on the floating catchment formation of Ho Chi Minh City’s alleyways. Focusing on method (FCA), which methodologically enables the the efficiency and robustness in ant networks of gal- integration of multiple sources of information into a leries, they state that the efficiency of self-organized single one. KC, Corcoran, and Chhetri (2020) employ graphs is reached by increasing the meshedness, that the enhanced two-step floating catchment method is by merging trees. However, this needs further inves- (E2SFCA) to investigate the optimality of fire station tigation to be applied in an urban context, especially locations in relation to changing spatial distribution of combined with narrow street widths. Based on this the population. Xia et al. (2019) propose a model of study, the developed model analyzes the increase of spatio-temporal accessibility to emergency medical meschedness in the alleyway network and indicates services (ST-E2SFCA), considering temporal variation appropriate positions to add new links in the street in population distribution in the greater Tokyo area network. through a large volume of GPS data of millions of Spatial accessibility to emergency services is an users and compare the accessibility over space and important indicator for evaluating the effectiveness of time. public health services (Xia et al. 2019) and its enhance- Both studies on emergency service accessibility ment is a key strategy to help improve emergency consider temporal variation in population distribution response, minimize property loss, and reduce injuries on a different scale, long-term population growth and and deaths (KC, Corcoran, and Chhetri 2020). There real-time population location, as well as the location of have been numerous studies investigating accessibility emergency facilities. However, a large set of data is of emergency services based on network theory, con- necessary to apply these models on an urban area, sidering potential modifications in the network struc- which is difficult to procure in fast-developing cities, ture. The risk of disruptions in the network can have a where data availability is scarce, often not collected or great impact on the accessibility of emergency ser- difficult to obtain. Therefore, this study introduces an vices. Ertugay, Argyroudis, and Düzgün (2016) consider easily operated evaluation model, based solely on net- road closure probabilities in the case of an earthquake work data, such as the street network and street width, in their accessibility analysis, due to ground failure, which can easily be applied in almost in every urban damage to bridges and overpasses, and collapses of area without relying on large sets of data. buildings adjacent to road edges. Tan, Hu, and Lin In terms of evacuation safety, Brachman and (2015) developed an agent-based simulation of build- Dragicevic (2014) developed an emergency evacuation ing evacuation, considering evacuee’s spatial model based on network science accounting for 6 TRAN T. ET AL. biological variables (fear, survival instinct) and social level, combining pedestrian evacuation of residents variables (emergency management, disaster response) and the network characteristics and topology, stres- in addition to the physical variables (transportation sing the importance of the conservation of traditional network, location of evacuees). Because of its ability neighborhoods, especially in fast-developing cities, to examine various emergency scenarios, the study where community ties are deeply embedded in the requires a great number of datasets, such as GIS data, daily life and social structures of residents. In the case to determine the number and spatial location of evac- of Ho Chi Minh City, Gibert and Pham (2016) empha- uees, registered vehicle data to evaluate the traffic flow size the importance of the multifunctionality and and the location of hazards, such as oil refineries. adaptability of the alleyway neighborhoods from a Similarly, Chen et al. (2020) investigate the evacuation socio-economic point of view and their ability to wel- vulnerability in urban areas in a fine-grained spatio- come a multiplicity of activities and adapt to all social temporal scale, overlapping mobile phone location structures. datasets with the road network of Shanghai. While Although many studies have evaluated urban safety, these studies provide valuable concepts to assess the there seems to be no method that can be adapted in complexity of emergency evacuation through ample traditional and dense urban areas, especially focusing data availability, they do not propose a solution to on mesh and tree-shaped network topologies, com- enhance the current evacuation safety of the urban monly found in self-organized urban structures. structure. In this regard, this study evaluated the cur- Furthermore, research about Vietnamese cities and rent evacuation safety level and evaluates the effi - their urban structures from a computational approach ciency of network modifications, such as edge is scarce, which could be due to the difficulty of obtain- widening or adding new links, essential for the devel- ing sufficient information. Therefore, this study intro- opment of dense urban areas. duces an easily operated evaluation model, based In this context, Zuo et al. (2021) propose a method solely on network data; this is a necessity as Ho Chi of improving emergency evacuation effectiveness Minh City is experiencing rapid development and under the restriction of limited available land resources modernization. in high-density areas, which is combined with urban renewal planning, aiming at optimizing the layout of 3. Analysis target and data processing shelters and evacuation passageway. While the safety of the evacuation process is improved in a “micro The model is based on the geometrical structure of update” manner, they do not consider the addition of the street network in order to calculate the safety level new links in the street network, which is a highly of a position in the network, more precisely, the acces- efficient improvement method for mesh and tree- sibility of emergency services and the evacuation shaped networks, dealt with in this study. danger. Oki and Osaragi (2014) focus on wide-area evacua- tion difficulty in case of a major earthquake in a 3.1. Weighted alleyway neighborhood network densely built-up wooden residential area in Tokyo combining a multi-agent simulation model with a An alleyway neighborhood located in Ward 2 of property damage model describing building col- District 5 in Ho Chi Minh City is taken as a case study lapse. However, in the context of rapid urbanized area. The site has a total area of 0.127 km and contains areas, building information is difficult to obtain and a population density of 37,486 pers./km . (Ho Chi Minh building materials vary between regions and cul- City Statistics Office 2019) tures. Therefore, this study proposes a model that The input data representing land plots was created can be applied in less data covered areas, where based on the digital format map provided by the “Ho urban renewal and safety improvement is highly Chi Minh City Urban Planning Information” (Ho Chi needed. Minh City Department of Planning and Architecture). Zhang et al. (2021) research about pedestrian eva- The CAD drawing contains land plot borders enabling cuation modeling and simulation in multi-exit scenar- the measurement of street width attributed to the ios using a social force model, while Chan (2019) drawn edges along with their length. A new edge is stresses the importance of a mandatory building drawn when there is a change in direction or at an inspection scheme and sustainable building mainte- intersection as no curves are drawn in the network and nance for safer and healthier cities, highlighting the all edges represent straight lines. The entrances of present situation of building deterioration in Hong each house, hereinafter referred to as house nodes, Kong. In order to ensure the health and safety of are connected to the street network by an edge traced people, a safe evacuation process has to be ensured from the middle of the land plot border perpendicular in different emergency scenarios, reaching from the to the street and has a set length of 0. A node is architectural building scale to large urban areas. This positioned at each end of an edge. The edge attributes study focuses on evacuation safety on a neighborhood are classified into nine categories. The edges of the JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 7 Figure 3.1. Network with street width categories. main street representing the border of the alleyway are classified as narrow edges, and all edges with a neighborhood, house edges, and the remaining edges width above 3.5 m are classified as wide edges, are classified depending on the street width with a set including safety threshold of 3.5 m, as shown in Figure 3.1. those with the main street attribute. The safety level This dataset was extracted and applied in a Python of a house is equal to its adjacent edge. In the next program. The proposed model was built using step, all the narrow edges are removed from the net- NetworkX, an open-source Python package, by struc- work graph. Setting the intersection nodes between turing the extracted data creating an adjacent list with the alleyway network and the main street, hereinafter the neighboring nodes of each node and an attribute defined as evacuation points, as source nodes and all list containing information of the edge length and nodes in the network as target nodes, all nodes that street width category. can be reached belong to the accessible network, as referred hereinafter, and all others belong to the inac- cessible network. In 3.2. Accessible and inaccessible networks Figure 3.2, the blue links in the network represent the alleys where emergency vehicles can reach, and According to previous research by Lin and Chen (2009), the red links represent the narrow alleys that are the necessary space for fire operations can be esti- impassable for emergency vehicles or wide edges mated by adding the width of the fire engine, which that are not connected to the accessible network, is a maximum of 2.6 m, the firefighter operation space and therefore belong to the inaccessible network. The of 0.4 m, and the space for the switch door, which is 0.5 evacuation points are the target locations where resi- m. Therefore, the sum of 3.5 m should be the minimum dents should go in case of an evacuation. The nodes space for firefighting and is set as the safety threshold where the accessible network switches to the inacces- to analyze the accessibility of emergency vehicles in sible network are defined as accessible points, as they the network. First, all edges with a width below 3.5 m 8 TRAN T. ET AL. Figure 3.2. Accessible and inaccessible networks. represent the furthest point an emergency vehicle can adequate emergency response due to accessibility dif- enter inside the network. These nodes are key for ficulties of emergency vehicles delaying the response calculating and analyzing the accessibility and evacua- time of emergency services. tion risk inside the network in the next steps (See (2) The unreachability risk and (3) edge responsibil- Figure 3.3). ity consider the risk residents face on the evacuation The accessible points are the generators used to route, if a fire or accident were to cause a disruption in calculate the accessibility network voronoi diagram, the network. The unwell connected network and dead- assigning each node on the inaccessible network to end streets cannot offer two-way evacuation route the closest accessible point. The evacuation points are possibilities for residents, representing a great danger the generators used to calculate the evacuation net- in case of an emergency. work voronoi diagram, assigning each node in the net- (4) The flow capacity focuses on the evacuation work to the closest evacuation point as shown in process of residents if a building fire or the spread of Figure 3.4. smoke forces them to evacuate to a safe area. The narrow and labyrinthine network of the alleyways cre- ates a dangerous environment during the evacuation 4. Safety level evaluation model process for residents. While the accessibility risk and unreachability risk This section introduces four evaluation variables of the give us the safety information of a specific location in safety evaluation model, namely the (1) accessibility the network (node), identifying houses at risk, the edge risk, (2) unreachability risk, (3) edge responsibility, and responsibility and flow capacity focus on the passing of (4) flow capacity, as shown in Figure 4.1. a location in the network (edge), analyzing the robust- The four evaluation variables have been developed ness of the network. to react to specific risks criteria and emergency situa- Fire policy and mitigation strategies in developing tions, analyzed based on the available network data. countries are constrained by inadequate information, (1) The accessibility risk focuses on the response of which is mainly due to a lack of capacity and resources emergency services like an ambulance or firetruck to a for data collection, analysis, and modeling (Masoumi, location in the network, in the case of a person need- van L.Genderen, and Maleki 2019). In densely built ing medical care or a building fire. Dead-end streets urban areas, building fires represent a high risk as it and narrow street widths increase the risk by impeding JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 9 Figure 3.3. Accessible and evacuation points. Figure 3.4. Accessibility Network Voronoi diagram (left) and Evacuation Network Voronoi diagram (right). can easily spread to adjacent buildings. In addition, the the consequence of a disaster. First, the accessibility of access to rescue services in case of a medical emer- emergency vehicles, such as a firetruck or ambulance, gency can greatly affect health, injuries or deaths. which is fundamental to respond to the disaster. In this The extraction of the evaluation variables has been case, the location of accessible points is essential to developed from a bottom-up approach. Due to the evaluate the difficulties emergency responders might limited data availability, possible indicators have face in the network, leading to the development of the been derived from the network topology and street first evaluation variable, i.e., the accessibility risk. width in relation to probable emergency scenarios. As Second, the evacuation of residents is essential to indicated in Figure 4.2, two outcomes are defined as ensure their safety after a disaster occurs, thus the 10 TRAN T. ET AL. Figure 4.1. Structure of the safety evaluation model. Figure 4.2. Evaluation variables extraction process. importance of the location of evacuation points in the point, which represents the nearest location accessible network. In this context, residents can be exposed to by a vehicle from that respective node. While the ideal two possible dangers. On one side, the risk of a poten- accessibility model includes the optimization of the loca- tial disruption in the network as a consequence of the tion of emergency facilities and the time to reach each disaster, disconnecting houses from an evacuation house on a city-wide scale, this index focuses on the scale point and leading to the inability of residents to evac- of a local neighborhood to provide emergency response uate to a safe area. On this account, two evaluation services on a community level. This first step is crucial to variables are developed: the unreachability risk, evalu- achieve a safe city-wide access in the future. ating the potential disconnection of a house from an Accessibility risk of node i is given by evacuation point, and the edge responsibility, evaluat- shortest distance from i to the nearest ing the criticality of an edge, if that edge is removed in d ¼ accessible nodeði : inaccessible nodeÞ (1) the network. On the other side, the risk caused by 0ði : accessible nodeÞ people flows in the network, which can lead to con- gestion and delay in the evacuation process, is mea- The accessibility risk d is expected to be proportional to sured by the fourth evaluation variable, the flow the required emergency response time from node i to capacity. While the studied network characteristics the nearest accessible point; the higher the d value, the enhance these risks, the developed evaluation vari- higher the risk. In Figure 5.2 the distance to the nearest ables are able to detect the most vulnerable locations accessible point is represented by a colorbar that and support the improvement of the safety level of the depends on the value of d as shown in the legend of street network. the figure. The dark blue color represents nodes when d is equal to 0, which means houses are located on the accessible network. The color gradation from blue to 4.1. Accessibility risk red represents the increase in distance from a node to To get a grasp of the current safety level in the network in an accessible point, and therefore a higher risk as emer- terms of accessibility, all the house nodes were measured gency responders need more time and have more diffi - depending on the distance to the nearest accessible culties to reach that node in an emergency situation. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 11 Figure 5.2(c) illustrates edges with an edge respon- 4.2. Unreachability risk sibility r equal to or greater than 1, which are located Each house in the area represents a possible location of on tree-shaped parts of the network ending in a cul- a fire or accident that can cause a disconnection in the de-sac. This leads to a higher risk if the network were to network. To simulate all possible scenarios, any one of be interrupted in case of an emergency, i.e., there is no the edges is removed one after another in front of two-way evacuation possibility. The color gradation every house, breaking the connection in the network from green to red shows the increasing number of in that location. By calculating the paths from all house houses disconnected from any evacuation point if nodes to any evacuation point, the unreachability risk that edge is removed. The highest risk is represented u represents how many times a house node j is not in dark red and is located at the root of the tree net- connected to any evacuation point through all possi- work. In a tree topology, even a single point of failure ble scenarios. can disrupt the connection from a node to a safe area. Unreachability risk of house node jð¼ 1; 2; . . . ; nÞ is The hierarchy of the tree branches and the total num- given by ber of houses on the tree network reflect the respon- sibility an edge holds in case of disruption of that edge u ¼ δ ; j jl l¼1 in the network. where 0 1 if house node j is connected 4.4. Flow capacity < to any evacuation node when B C 0@ A: The flow capacity is calculated by simulating an eva- δ ¼ edge lð¼ 1; 2; . . . ; mÞ; adjacent (2) jl to any house; is removed: cuation process from each house in the network to an 1ðotherwiseÞ: evacuation point and is evaluated based on two para- meters, namely edge capacity and bottleneck risk. For n : Total number of house nodes this evaluation variable, the input data of the edges in m : Total number of edges adjacent to any house the network was modified by dividing the edge length If u is equal to 0, house node j has at least a two-way into equal segments of 1.0 m. This index measures the evacuation route to any evacuation point, represented density of evacuees at the peak time of edges and in dark blue in Figure 5.2(b). If u is equal to or greater detects the locations of bottleneck risks, delaying the than 1, house node j is located on a tree-shaped net- evacuation process. work, inside the mesh network of alleyways. The gra- dation of colors from blue-green to red shows the 4.4.1. Edge capacity increase in risk of unreachability as shown in the The edge capacity represents the average pedestrian graph legend. space on an edge at peak time during the evacuation process and is measured by dividing the edge area by the maximum number of residents on that edge dur- 4.3. Edge responsibility ing the evacuation process. Evacuees reach the nearest The edge responsibility r of edge k is defined by the evacuation point using the shortest path. number house nodes that are disconnected from any Edge capacity of edge k is given by evacuation point in all scenarios simulating a discon- c ¼ (4) nection in the network by removing edge k. k maxfP g kt The edge responsibility represents the importance of an edge to provide an evacuation route between a A : Area of edge k house and a safe area. P : number of evacuees on edge k in time t kt Edge responsibility of edge k is given by We have three assumptions for the calculation of P : evacuees are distributed equally among all house r ¼ h ; kt k kj j¼1 (1) nodes , they have the same walking speed and the same evacuation starting time. Based on these where assumptions, we can calculate P by the distance kt if house node j is connected between any house node and edge k. Max t{P } repre- < kt 0 : to any evacuation node when h ¼ (3) sents the maximum number of evacuees on edge k kj edge k is removed during the whole evacuation process. If max t{P } is kt 1ðotherwiseÞ: equal to 0, c is not defined. The edge capacity c k k n : Total number of house nodes represents the danger of congestion; which is The average number of residents per household has been taken as 4.7 based on the population density indicated for District 5 in the Ho Chi Minh City Statistical Yearbook 2018 (Ho Chi Minh City Statistics Office 2019) 12 TRAN T. ET AL. especially high if two evacuation routes merge without the bottleneck risk. The highest risk of flow capacity an increase of street width, i.e., increase of available arises when the area per person of the edge capacity is area per evacuee. In Figure 5.2(d), c is illustrated by a low, as shown in dark blue, and the difference in gradation from dark blue to light blue. The darker the density of the bottleneck risk is high, as shown in blue, the less space per evacuee is available, which dark red. represents a higher risk; the lighter blue means the available space per evacuee on that edge is larger, 5. Model behavior and application on typical which is safer. street network topologies The evaluation model is first applied on a typical street 4.4.2. Bottleneck risk network topologies to observe the behavior of the The bottleneck risk of a node is caused by the decrease in developed model and to investigate how the model area per person between its adjacent edges in the direc- recognizes locations at risk in different network topol- tion of the evacuation path, which creates a bottleneck ogies. Three abstracted paradigms of street networks effect, and can cause delays in the evacuation process. were constructed to demonstrate how the developed Bottleneck risk of edge k is given by evaluation model analyzes distinctive patterns. The b ¼ c c (5) k k k street network paradigms represent an abstract form of the main categories of typical street patterns found C : edge capacity of edge k s adjacent edge in in cities across regions and cultures. the direction of the evacuation path For each paradigm, first the virtual street network is If b is negative, edge k has no bottleneck risk; if b is drawn inside a square measuring 300 × 300 m. Then, the k k equal to or superior to 0, the end node of edge k in the perimeter of the network is classified as main street, direction of the evacuation route has a bottleneck risk. which belongs to the accessible network. In the next As seen in Figure 5.2(d), the bottleneck risk is repre- step, the accessible points are first located on the peri- sented by a gradation of red and proportionate size of meter, hereinafter referred to as case (1). Then, the acces- the dot. The darker and bigger the red dot, the higher sible points are located inside the network, expanding Figure 5.1. Typical street network topologies. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 13 Figure 5.2. Application of the evaluation model on a virtual street network. the accessible network inside the network, referred to as Figure 5.2 (a) shows the accessibility risk for all case (2). Finally, the other street width categories are topologies. In the case ofgrid (1), the distance to an distributed randomly inside the network in both cases. accessible point is increasing continuously towards the As seen in Figure 5.1, the first street network para- center of the network due to the equal distance dis- digm is a regular grid network and represents a recti- tribution between nodes on the grid network. linear network with a pattern of streets running at right Regarding grid (2), the distance to an accessible point angles. The second street paradigm is a degree-3 net- reacts to the location of the accessible points inside work, also referred to as T-type network. This topology the network, but d remains less than 75 m. The acces- has a high proportion of degree 3 nodes. The third sibility risk shows a similar result in the degree-3 net- street paradigm is a tree and dendritic network. work, where the center area of the network is the most 14 TRAN T. ET AL. vulnerable in case (1) and reacts to the expansion of grid structure is predominant in the center and the accessible network in case (2). Regarding tree and becomes less dense towards the outskirts of the dendritic networks, the accessibility risk increases with city. Narrow and sinuous alleyways composed of the total length and total number of houses of the tree tree and mesh networks fill up the spaces of var- network and is the highest at the “leaves” of the tree ious scales between the grid structure. network. This behavior can be observed in both cases, but the risk decreases in case (2). Naturally, the acces- 6.1. Analysis of the current situation sibility risk is lower in case (2) of all topologies due to the extension of the accessible network. The four evaluation variables have been applied to the In terms of unreachability risk and edge responsi- network of the analytical area in Ward 2, District 5 of bility, the evaluation model reports that all u and r are Ho Chi Minh City. The analyses of the network give a j k equal to 0 in the entire network for the grid and clear understanding of the current level of safety and degree-3 topologies due to their high connectivity. In helps to identify the most vulnerable areas in the net- the case of tree and dendritic networks, the unreach- work for each criterion. ability risk is the highest at the “leaves” of the tree The analysis of the Accessibility risk shows that networks and increases with the size of the tree net- currently 74.12% of the houses in the network are work (Figure 5.2 (b), while the edge responsibility is the not accessible by an emergency vehicle, among highest at network roots, connecting a tree network to which 38.33% are located further than 50 m from an the accessible network as seen in Figure 5.2(c). As seen accessible point (Figure 6.2). As can be observed in in Figure 5.2 (d), the application of the flow capacity Figure 6.3, houses located on a tree-shaped network evaluation on all network topologies shows that the are affected by the unreachability risk as they are not result does not depend on the network topology, but provided with a two-way evacuation route. Currently, rather on the distribution and location of segments 34.87% of the houses are at risk of being unable to with different street widths in the network. reach an evacuation point if an accident were to inter- rupt passage through the network. The Unreachability risk increases proportionately with the size of the tree 6. Model application to Ho Chi Minh City network and the number of houses located between a This section introduces the application of the model in house and an evacuation point. a real urban area, an alleyway neighborhood in Ho Chi In terms of edge responsibility, the risk occurs in tree- shaped networks providing the only access between Minh City. In section 6.1, the current safety level of the houses and an evacuation point. The risk level rises street network is analyzed, then improvement strate- gies are applied in three case studies in section 6.2. with an increase in the length and hierarchy layers of the tree-shaped network as well as with the number of First, the methodology, which combines edge widen- houses located on that network. Consequently, the ing and adding new edges in the network, is explained in section 6.2.1, then the strategies to choose the edge responsibility varies in accordance with the hier- archy of the branches in the tree network and the location of edges for improvement are detailed in sec- number of attached houses and increases towards the tion 6.2.2. Finally, the improvement intervention results are described in section 6.2.3, and compared root of the tree network. Here, 21.59% of the current network has a tree shape, creating subnetworks with and discussed in section 6.3, as illustrated in Figure 6.1. branches of various sizes as seen in Figure 6.4. Ho Chi Minh City was chosen as the analytical area for this study; it exhibits a very unique urban As for the flow capacity, illustrated in Figure 6.5, the trajectory after experiencing colonization, decades edge capacity is represented in blue and the bottle- neck risk of nodes is represented in red. The bottleneck of war, socialism, and de-urbanization, followed by risk of nodes with a higher difference in density the national reunification of 1976 and Đổi Mới reforms (Thrift and Forbes 1986). Before the two appears predominantly at intersections, while the edge capacity is influenced by the width of the street cities were merged and renamed Ho Chi Minh City and the number of times an edge appears on an after the reunification of the country, the French laid the structure for Saigon’s urban grid, while evacuation route and is scattered in the network. Cholon’s economic and social structure shaped its urban form. By 1955, a migration movement to 6.2. Case studies of improvement the city created serious housing problems and overcrowding and led to the spontaneous urban 6.2.1. Methodology development, starting the formation of Ho Chi Three case studies were conducted using the simulation Minh City’s typical alleyway system, which has model focusing on improving the safety level in the net- lasted until today. After the war, the migration work by combing two strategies. The first strategy movement continued and led to further densifica - involves adding a new edge in the network. This method tion of the alleyway system. Therefore, the primary changes the network geometry and has a high impact on JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 15 Figure 6.1. Structure of section 6.Model application to Ho Chi Minh City. the shortest path routes and two-way evacuation routes In the case studies, both methods were combined, by linking dead-end streets and creating new loops inside aiming to improve the safety as much as possible by the network. The addition of a new edge in the network keeping the area subject to change, thereby keeping can be highly efficient while having a small impact on the the costs and impact on the local community as low local community. The second strategy involves widening as possible. The total length of added edges in the existing alleyways. This method has a direct effect on the network and the total area to expand for edge accessibility of emergency vehicles by expanding the widening was equal in all the case studies. The accessible network further and also increases the area total area for edge widening was set at 250 m , of edges on the evacuation route. and the total added edge length was set at 64 m 16 TRAN T. ET AL. Figure 6.2. Accessibility risk. Figure 6.3. Unreachability risk. with a width of 3.5 m. For each case study, the area 6.2.2. Strategy for safety level improvement of improvement was gradually expanded and was The choice of edges to add or widen was decided referred to as high concentration intervention area based on different strategies. Three strategies were in case study 1, medium concentration intervention applied for widening edges. area in case study 2, and low concentration interven- tion area in case study 3, as seen in Fig.5.5. (a) The first strategy to select edges to widen Subsequently, the concentration of interventions of involves connecting edges on the inaccessible improvement decreased from case study 1 to 3. network with a street width greater than 3.5 m JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 17 Figure 6.4. Edge responsibility. Figure 6.5. Flow capacity. to the accessible network (see Fig. 2.1). This (c) The third strategy involves connecting two approach is represented in red in Figure 6.6. accessible networks to one another (see (b) The second approach involves extending the Figure 3.2). This approach is represented in accessible network further into the network yellow in Fig.6.6 and has been combined (see Fig. 2.2). This approach is represented in with the addition of new edges as demon- orange in Figure 6.6. In case study 1, segment strated in case study 2 with the new edges (3) combines this approach with the addition of [4] and [5] and in case study 3 with the new the new edges [2], [3] and [4]. edges [2] and [3]. 18 TRAN T. ET AL. The locations for adding new edges and creating new links in the network targets tree networks with the highest risks first (See Figure 6.2 to 6.5). After targeting the most unsafe areas in the network, the focus of adding new edges shifts to medium and small size networks. The biggest tree networks are connected to the mesh network with the new edge [5] in case study 1, the new edges [3] and [4] in case study 2, and the new edge [2] for the tree network with the highest risk and the new edges [4] and [5] for the second most vulnerable tree network in case study 3. 6.2.3. Improvement intervention results The safety enhancement of the improvement interven- tions varies for each case and is described as follows: Compared to the current situation, the accessibility risk improves considerably inside the intervention area in case study 1 as the accessible network is expanded by both, widening edges and adding new edges con- densed in a small area as seen in Figure 6.7. While widening edges have no influence, adding new edges has similar impacts on the unreachability risk and edge responsibility. No high-risk areas remain in any of the case studies, and the total length of tree networks decreases continuously compared to the cur- rent network from case study 1 to 3 as seen in Figure 6.8 and Figure 6.9. Widening edges can improve the edge capacity, and new links can alter the evacuation route, but not all interventions have an impact on the flow capacity. Adding a new edge only has an impact on the flow capacity if this new connection in the network modifies the shortest path between any house node and an evacuation point. For instance, the new link [2] in case study 1, seen as having an edge capacity value in blue in Figure 6.10, not only decreases the length of the evacuation route of seven houses but also modifies the bottleneck risk in that area: in this case, at the intersection of segment (3). Widening edges can improve the edge capacity of that edge and improve or remove the bottleneck risk around that edge. The new edge [6] in case study 2 reduces the bottleneck risk in proximity of that edge but increases the bottle- neck risk in other locations nearby. This is affected by the change in evacuation routes of 11 houses passing through the new link, which shortens the distance between these houses and the nearest evacuation point. In case study 3, the widening of edges in seg- ment (1) removes a bottleneck with a high-risk value, but smaller bottleneck risk locations appear on the intersection of segment (2) (see Figure 6.10). 6.3. Discussion After applying the proposed model on three case Figure 6.6. Added and widened edges in case studies 1, 2, studies, an accumulative database of graphs and and 3. charts was created to compare the results of the JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 19 Figure 6.7. Accessibility risk. Figure 6.8. Unreachability risk. Figure 6.9. Edge responsibility. Figure 6.10. Flow capacity. case studies with the current situation. As seen in studies, respectively. In case study 3, the houses Figure 6.11, compared to the current situation, the located further than 80 m from an accessible point houses in the network that are inaccessible for an declined sharply from 15.98% in the current situation emergency vehicle decreased by 15.87%, 19.58% to 1.68% (Figure 6.12). While the number of accessible and, 13.89% in the first, second, and third case houses increased in all case studies, from the 20 TRAN T. ET AL. Figure 6.11. Comparison of accessible and inaccessible houses in all case studies. Figure 6.12. Distance distribution between inaccessible houses and accessible points for all case studies. perspective of accessibility for emergency vehicles, 51.47% compared to the current situation, leaving the best improvement was observed in case study 2, only 11.56% of the total network length tree-shaped. with an augmentation of 52.65% of the accessible Figure 6.16 shows that case study 3 has the lowest houses compared to the current situation. edge responsibility risk for each value of r . The risk of unreachability decreased continuously Regarding flow capacity, no case study showed a from case study 1 to 3, with a reduction in risk of clear improvement in both parameters, edge capacity, 35.43%, 44.37%, and 52.32% in the first, second and and bottleneck risk. In terms of third case studies, respectively, compared to the cur- case study 2, the number of edges with c less than rent situation (see Figure 6.13). Case study 3 has only 0.15 m increased by 48.6% compared to the current 16.82% of the houses with no two-way evacuation, network; however, it also showed the highest improve- compared to 34.87% in the current situation, and as ment with an increase of 24.29% in the safest category seen in Figure 6.14, exhibits the most improvement for with c greater than 0.6 m (Figure 6.17). The total each value of u . number of bottleneck risks decreased by 7.19% in The addition of new links in the network enabled case study 1, 10.07% in case study 2, and 9.35% in the connection of tree subnetworks to the mesh net- case study 3 compared to the current network as work, thereby providing a two-way evacuation possi- seen in Figure 6.18. bility in these areas. All case studies reduced the edge Figure 6.19 shows the social impact of the case studies responsibility in all risk levels. As seen in Figure 6.15, on the local residents. While alley widening impacts case study 3 showed the strongest improvement by households by reducing the area of their land plot, add- reducing the total length of tree subnetworks by ing new edges might force residents to relocate. While JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 21 Figure 6.13. Comparison of the frequency of houses and the unreachability risk for all case studies. Figure 6.14. Frequency of the level of unreachability risk among houses with no 2-way evacuation route in all case studies. Figure 6.15. Comparison of length of the network and risk of edge responsibility in all case studies. 22 TRAN T. ET AL. Figure 6.16. Comparison of the network length and the edge responsibility risk categories on tree-shaped networks. Figure 6.17. Comparison of the number of edges and the edge capacity. Figure 6.18. Frequency of the bottleneck risks. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 23 Figure 6.19. Comparison of the social impact in all case studies. the number of households forced to relocate were similar gradually with an increase in the number of houses in all case studies, the number of households affected by located between a house and an evacuation point on a alley widening was the highest in case study 2 and the tree-shaped network. In terms of edge responsibility, lowest in case study 3. Case study 3 applied on a low the risk is the highest at the roots of tree networks, and concentration area had the lowest social impact, while there is a parallel increase with the total length of the showing the best improvement for the accessibility risk, tree network and the number of houses located on especially in the high-risk categories and had the lowest that tree network. Regarding flow capacity, the edge unreachability risk and edge responsibility among all case capacity depends on the distribution of street widths studies conducted in this paper. across the network, and the occurrence of bottleneck If the weakest point in the network is targeted for alley risks is higher at intersections. improvement and is chosen as an investment project, it As confirmed in the result of the case studies, the provides the most efficient result of network-wide safety effectiveness of the improvement interventions varies improvement for the accessibility risk, unreachability risk, depending on the evaluation variables and the and edge responsibility, but this type of scenario is not method applied: edge widening or adding an edge. valid for the concept of flow capacity, where modifica - Regarding adding new edges in the network, the tions in the network only have a local result. However, the accessibility risk can only be improved if the added flow capacity can detect which intersections in the net- edge is connected to the accessible network, but if work are actual intersections on the evacuation routes the added edge modifies the evacuation route of resi- and therefore have a bottleneck risk, which cannot be dents, it can improve or deteriorate the flow capacity, derived from the network topology. In Figure 6.5, the depending on whether the evacuees using the short- bottleneck risk appears to be especially high at intersec- est path from a house to the nearest evacuation point tions, but this is not the case for intersections located at are led through more or less crowded edges, as it the border of the Voronoi network cells where residents might increase the duration of the evacuation process. evacuate in opposite directions. In terms of unreachability risk and edge responsibility, adding a new edge transforms a tree network into a mesh network and provides a two-way evacuation, 7. Conclusion which increases the safety the most. An original methodology for evaluating accessibility As for widening edges in the network, with the and evacuation safety using network analysis was cre- applied strategy, the accessibility risk improves when ated and applied in Ho Chi Minh City’s alleyways. the accessible network is extended, but it has no effect The evaluation of the current street network and the on the unreachability risk and edge responsibility. modifications in the case studies disclosed various While improvement interventions have an impact on results and features of the evaluation variables and safety improvement over a wider area in the network improvement methods. Showing the highest risk at for other evaluation variables, in terms of flow capacity, “leaves” of inaccessible tree-shaped networks, the widening edges only has a local effect in the network. methods applied to improve the accessibility risk The prospective intervention projects were com- involve a combination of reducing the maximum dis- pared in terms of safety improvement, economic tance to an accessible point and extending the acces- impact (defined by the area to modify), and social sible network. The Unreachability risk worsens impact (defined by the number of households forced 24 TRAN T. ET AL. to relocate and households having to give away a part Disclosure statement of their land plot). This evaluation model is not only No potential conflict of interest was reported by the author(s). useful to analyze the current risk level in the network but is also a powerful tool for urban planners and governments to manage and estimate the efficiency Funding of future infrastructure improvement projects. This work was supported by JSPS KAKENHI Grant Number JP From an academic point of view, this research is 20H02327. based on existing network analysis models, such as the Network Voronoi algorithm, which already ORCID proved their efficiency in previous researches but evolves by introducing a novel method adapted to Tran Thi To Uyen M.N http://orcid.org/0000-0002-9218- a specific network topology and considering the urban context and narrow street widths. In prac- tice, the structure and safety of self-organized set- References tlements seems complex, and dangerous locations cannot be pointed out intuitively or are often Brachman, M. L., and Dragicevic S. 2014. “A Spatially Explicit overlooked in traditional urban planning Network Science Model for Emergency Evacuations in an Urban Context.” Computers, Environment and Urban Systems approaches. In this regard, the developed model 44: 15–26. doi:10.1016/j.compenvurbsys.2013.10.009. is able to identify vulnerable locations based on Buhl, J., J. Gautrais, N. Reeves, R. V. Solé, S. Valverde, P. Kuntz, network data solely and evaluate future improve- and G. Theraulaz. 2006. “Topological Patterns in Street ments, which is very helpful for urban manage- Networks of Self-organized Urban Settlements.” The ment and sustainable development, especially in European Physical Journal B - Condensed Matter and fast developing cities. Complex Systems 49 (4): 513–522. doi:10.1140/epjb/ e2006-00085-1. This research, however, is subject to several limita- Buhl J., Gautrais J., Solé R.V., Kuntz P., Valverde S., tions that could be addressed in future research. As Deneubourg J.L. & Theraulaz G.2004. “Efficiency and mentioned in section 3.1, in terms of accessibility, one Robustness in Ant Networks of Galleries.” The European limitation of this study is the boundary of the area, Physical Journal B 42 (1): 123–129. doi:10.1140/epjb/ meaning that activities or the location of emergency e2004-00364-9. Chan, D. W. M. 2019. “Sustainable Building Maintenance facilities outside the analytical area, are not considered for Safer and Healthier Cities: Effective Strategies for in the model. Moreover, this analysis is based on build- Implementing the Mandatory Building Inspection ing fires and medical emergencies and does not include Scheme (MBIS) in Hong Kong.” - Journal of Building other natural or human-made disasters that can occur in Engineering 24: 100737. doi:10.1016/j.jobe.2019.100737. dense urban areas. Another limitation is, that this Chan, H., and Akoglu, L. 2016. “Optimizing Network research is based on network analysis, more specifically Robustness by Edge Rewiring: A General Framework.” Data Mining and Knowledge Discovery 30 (5): 1395–1425. the network Voronoi algorithm, which is based on the doi:10.1007/s10618-015-0447-5. assumption of the shortest path rule. Moreover, this Chen, J., T. Pei, M. Li, C. Song, T. Ma, F. Lu, and S. Shaw. 2020. assumption is based on a number of other assumptions, “An Enhanced Model for Evacuation Vulnerability for example, that all residents are familiar with the Assessment in Urban Areas.” Computers, Environment and whole network structure and positions of locations. Urban Systems 84: 101540. doi:10.1016/j. compenvurbsys.2020.101540. Furthermore, it simplifies the complexity of human deci- Chen, P., H. Wu, and Y. Hsu. 2019. “Widening Narrow Alleys to sions made in case of an emergency, which can be Enhance Response Efficiency for Fire Emergency from the unrealistic in certain situations. Finally, the safety thresh- Perspective of Urban Roadway Network Analysis .” Journal old for the access of emergency vehicles, set at 3.5 m, is of the Eastern Asia Society for Transportation Studies 13 only based on the street width and does not consider 2598–2613 . the geometry of the space such as corners and turning Chen, Jiayan, Jia Yu, Jiahong Wen, Chuanrong Zhang, Zhan’e Yin, Jianping Wu, and Shenjun Yao. 2019. “Pre-evacuation space for emergency vehicles. Time Estimation Based Emergency Evacuation Simulation The next step would be to expand the model to a in Urban Residential Communities.” International Journal city-wide scale and include the locations of emergency of Environmental Research and Public Health 16 (23): 4599. facilities to evaluate the accessibility and response doi:10.3390/ijerph16234599. time of rescuers more precisely. Furthermore, in Chan H., Akoglu L. ,and Tong H.2014. “Make It or Break It: terms of edge responsibility, the extent of overlapping Manipulating Robustness in Large Networks.” In Proceedings of the 2014 SIAM International Conference on parts of two-way evacuation routes should be verified. Data Mining, SDM’14 April 24-26, 2014 Philadelphia, Finally, the current model analyzes the four evaluation Pennsylvania, USA, pp. 325–333 variables independently, which should be combined in Ertugay, K., Argyroudis, S., and Düzgün, H. Ş. 2016. future research to grasp the full potential of improve- “Accessibility Modeling in Earthquake Case considering ment interventions. Road Closure Probabilities: A Case Study of Health and JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 25 Shelter Service Accessibility in Thessaloniki, Greece.” Oki, Takuya, and Toshihiro Osaragi. 2014. “A Method for International Journal of Disaster Risk Reduction 17: 49–66. Evaluation of Wide-area Evacuation Difficulty in case of A doi:10.1016/j.ijdrr.2016.03.005. Major Earthquake.” 12th International Conference on Gibert Marie and Pham Thai Son (2016) “Understanding Design and Decision Support Systems in Architecture and the Vietnamese Urban Fabric from the Inside”. Urban Planning, DDSS 2014 25-27 August 2014 Eindhoven, Producing and living the city in Vietnam, The Focus, Netherlands The Newsletter, No. 73. Tan, L., M. Hu, and Lin, H. 2015. “Agent-based Simulation of Ho Chi Minh City Department of Planning and Architecture, Building Evacuation: Combining Human Behavior with Ho Chi Minh City Urban Planning Information. Application Predictable Spatial Accessibility in a Fire Emergency.” for webpage and mobile devices. Accessed 4 11 2019 Information Sciences 295: 53–66. doi:10.1016/j. https://thongtinquyhoach.hochiminhcity.gov.vn/ ins.2014.09.029. Ho Chi Minh City Statistics Office. 2019. “Ho Chi Minh City Thrift, N., and D. Forbes. 1986. The Price of War: Urbanization Statistical Yearbook 2018”. Population, Labor and employ- in Vietnam, 1954-1985. London: Allen and Unwin. ment in 2018, table 12, Area, population and population Wang, Huijuan, and Piet Van Mieghem. 2008. “Algebraic density in 2018 by district. Connectivity Optimization via Link Addition.” In Johansson, E. 2006. “Influence of Urban Geometry on Proceedings of the 3rd International Conference on Bio- Outdoor Thermal Comfort in A Hot Dry Climate: A Study Inspired Models of Network, Information and Computing in Fez, Morocco.” Building and Environment 41 (10, Systems 25-28 November, 2008 Hyogo, Japan (22). ICST October): 1326–1338. doi:10.1016/j.buildenv.2005.05.022. (Institute for Computer Sciences, Social-Informatics and KC, K., J. Corcoran, and Chhetri, P. 2020. “Measuring the Telecommunications Engineering). Spatial Accessibility to Fire Stations Using Enhanced World Bank (2011) “Vietnam Urbanization Review, Technical Floating Catchment Method.” - Socio-Economic Planning Assistance Report.” In Chapter 3: Urban Expansion and Sciences 69: 100673. doi:10.1016/j.seps.2018.11.010. Spatial Development in Vietnam’s Cities, 3.1 Neighborhood Kitahara, Tomoji. 2001. “Conserving the Walkable Environment Urban Upgrading in HCMC, p.124. in the Neighborhood: A Case Study of the Improvement Xia, T., Song, X., Zhang, H., Song, X., Kanasugi, H., & Effect in Kyojima, Tokyo.” Australia: Walking the 21st Century, Shibasaki, R. 2019. “Measuring Spatio-temporal International Conference, 2001, Perth, Western Australia, 2 Accessibility to Emergency Medical Services through Big Lin, Chen. 2009. “A Study of Fire-fighting Strategy on Narrow GPS Data.” Health & Place 56: 53–62. doi:10.1016/j. Alleys Structures.” Journal of Disaster Mitigation and Rescue healthplace.2019.01.012. (Taiwan) 10 (1): 107–125. Zhang, D., Huang, G., Ji, C., Liu, H., & Tang, Y. 2021. Masoumi, Z., van L.Genderen, J., & Maleki, J. 2019. “Fire Risk “Pedestrian Evacuation Modeling and Simulation in Assessment in Dense Urban Areas Using Information Multi-exit Scenarios.” Physica A 582: 126272. doi:10.1016/ Fusion Techniques.” ISPRS International Journal of Geo- j.physa.2021.126272. Information 8 (12): 579. doi:10.3390/ijgi8120579. Zuo, J., J. Shi, C. Li, T. Mu, Y. Zeng and Dong, J. 2021. Novak D.C., and J.L. Sullivan. 2014. “A Link-focused “Simulation and Optimization of Pedestrian Evacuation in Methodology for Evaluating Accessibility to Emergency High-density Urban Areas for Effectiveness Improvement.” Services.” Decision Support Systems 57 (January): 309–319. - Environmental Impact Assessment Review 87: 106521. doi:10.1016/j.dss.2013.09.015. doi:10.1016/j.eiar.2020.106521.

Journal

Journal of Asian Architecture and Building Engineering – Taylor & Francis

Published: Mar 4, 2023

Keywords: Alleyway neighborhoods; Ho Chi Minh city; network analysis; urban safety

There are no references for this article.