Evaluation of Resilience in Historic Urban Areas by Combining Multi-Criteria Decision-Making System and GIS, with Sustainability and Regeneration Approach: The Case Study of Tehran (IRAN)

Abstract

Historic urban areas are the beating heart of the city, but neglecting them can lead to low resilience. Therefore, paying attention to their regeneration can create a sustainable city. The purpose of this study was to determine the resilience of neighborhoods in Tehran and evaluate effective criteria for the resilience increase. In this study, to evaluate the resilience of Tehran, initially, 18 criteria were considered. Then, using the Delphi technique, 14 criteria among them were selected for final analysis. Using the AHP multi-criteria decision-making method, the importance of each criterion was determined. Using GIS capabilities, the parameters map was prepared, and by combining the prepared maps with AHP weights, a resilience map was created. Finally, 20 neighborhoods with the lowest resilience were identified as priorities for stabilization and regeneration measures, and the criteria status used in them was examined. Results showed that deteriorated urban areas (19.53%) and construction materials (18.51%) were the most important criteria. Non-resilience areas were generally in the southern half of the city. 78% of 20 selected neighborhoods had deteriorated urban areas, while only 14% of the city deteriorated. Finally, by examining the criteria in neighborhoods with the lowest resilience, suggestions were made to regeneration, sustainability, and increase the resilience of these neighborhoods.

1. Introduction

Nowadays, urban resilience has become one of the most important topics due to the center of human accumulation in the city and the impact of resilience on urban areas [1,2,3,4]. The concept of resilience quickly found its place in the urban sustainability literature [5]. Urban resilience is often associated with the terms of sustainability, smart city, regeneration, etc. [6,7].

The word “resilience” is etymologically derived from the Latin word resilience, which means “to return, recoil” [8]. Resilience in the dictionary is the ability to return to the original state after a stressful and unusual situation [9]. The concept of resilience originates from ecology and refers to “a measure of the durability of systems and their ability to maintain stability in the face of change and turbulence and to maintain the same relationship between population or state variables” [10,11].

In recent scientific studies, ideas such as resilient communities, resilient livelihoods, and resilient community building have been widely used [12]. Since resilience is used in different fields, it has its own definitions in each scientific field. Some of these definitions include the definition in the environment [13], social sciences and the city [2], agricultural sciences and biology [14], engineering [15], and management [16]. However, one of the oldest and most complete definitions of resilience is to refer to resilience as “the ability of systems and their components to respond to internal or external disturbances which, after a recovery period, retain their essential characteristics” [17,18]. As Walker et al. stated [19], a resilient system is defined by two major characteristics: its ability to absorb changes and difficulties, as well as its ability to evolve, adjust, and endure while maintaining its fundamental structure and functions.

Urban regeneration is a comprehensive and integrated approach to solving urban problems in the region that is the goal of the operation, which ultimately leads to sustainable economic, physical, social, and environmental development [20]. It can be said that regeneration is a stage after determining resilience and in order to increase the resilience of areas with low resilience. Increasing attention to urban sustainability is affecting urban regeneration policy and performance [21]. Urban regeneration has been a sustainable theme in globalization and urbanization from developing to developed countries [22,23,24]. Nowadays, the term “urban regeneration” as a general term; includes other concepts such as revitalization, rehabilitation, reconstruction, and empowerment [25]. Social organizations use a combination of economic, environmental, cultural, and social activities to regenerate their communities sustainably [26]. Urban regeneration in traditional environments transforms the meanings embedded in social and cultural environments in these places [27].

The historic urban area is one of the main components of the urban system that takes care of social, economic, housing, transportation, security, and leisure needs, which, if not addressed, causes wear and inefficiency of this sector [28]. Historic urban areas, such as other types of cultural heritage, have significant values [29]. Built heritage has not only architectural and cultural value of construction but also cultural and anthropological cultural activities with their own characteristics [30,31]. The population of cities in developing countries is increasing every day [32], which makes it difficult to provide services to them [33,34]. Additionally, it causes the formation and continuation of an area of the city as a deteriorated urban area [35,36,37] which hinders a dynamic and continuous process of development in the city [38].

Disaster management and efforts to reduce risk in the historical areas are important for two reasons: (1) the value of historical heritage (2) and their economic and social role [39]. Damage causes inefficiency and physical deterioration of the city, which is a negative point [40,41]. In addition, historical regions have their own dynamics [42,43], so the revitalization of historical areas leads to the socio-economic development of these areas [29,44]. Residents of these buildings are from low-income and society disadvantaged groups who usually do not receive sufficient services and attention after a tragic event such as an earthquake [45].

Numerous studies have been conducted on urban resilience. Mahmodinia et al. [46] evaluated the resilience of the historical context of Yazd. They were using AHP measurement and layer composition in GIS to determine that the western and southern parts of Yazd city are in a bad situation of resilience. Rostami et al. [47] prepared resilience maps in GIS software by using Kendall’s W Ranks test in SPSS software and assigning appropriate weight to each of the indicators and has been analyzed. The results showed that the indicators of grading, age of buildings, quality of buildings, building structure, and permeability are effective in reducing the resilience of this area. In another study, Moaddab and Amini Hosseini [48] examined the indicators affecting the resilience of the historical parts of the city (with emphasis on old markets). By using the AHP method, it showed that: (1) The use of the building; (2) economic value of property and goods inside the shops; and (3) adaptability and dynamism of different groups for post-earthquake recovery are three important criteria in the resilience of old markets during an earthquake.

Historic urban areas are social-ecological systems, which are complex and adaptive to changing conditions. Throughout their evolution process, just like cities of which they are part, they can adapt and transform themselves without losing their essence, and after stages of crisis—a destructive incidence, such as a natural disaster, a war, an invasion, etc., they come back to an equilibrium, albeit a new one, which allows them to bounce back again. This evokes the concept of the adaptive cycle, which is a model with a focus of attention upon processes of destruction and reorganization and a crucial concept under the resilience theory. However, beyond the inherent scope and qualities of resilience, historic urban quarters are affected by specific vulnerabilities, in terms of different types and intensities of obsolescence, among which functional and locational ones can be considered as the most prominent, due to their negative impact on the performance of the area. Considering their inherent qualities as well as the vulnerabilities that historic urban quarters are faced with, their revitalization can be defined as an adaptation and transformation process. This brings the area to a new equilibrium, where the place assets are protected and the deterioration and decay are mitigated.

The concept of resilience is becoming increasingly important despite the many challenges that disrupt its implementation [49]. Nowadays, natural disasters and humanitarian actions are considered unpredictable urban threats [50]. Thus, improving the conditions of urban resilience infrastructure helps to recover the impacts inherent to disasters [50,51,52]. Urban resilience is a broad concept that encompasses all aspects of the city. The most fundamental issue in the discussion of resilience is the issue of regeneration of non-resilient areas. In this regard, considering that it is effective to determine the resilience of several criteria, it is necessary to use a multi-criteria decision-making system to determine the priority of each of the criteria. Finally, combining the results of this section with spatial information layers using the capability of GIS can provide a complete ranking of the resilience of urban areas and neighborhoods for regeneration. This research seeks to answer some basic questions: (1) What parameters are effective in determining the resilience of Tehran? (2) How important is each of the effective criteria in determining the resilient and non-resilient neighborhoods of Tehran? (3) Are the historical areas of the city resilient? (4) What is the status of effective criteria in resilient and non-resilient areas? Furthermore, according to the study of resilience in the historical context of Iranian cities, it was found that the historical areas of the cities in Iran are often non-resilient, so the hypothesis is formed that the historical areas of Tehran are not resilient. Therefore, this research seeks to answer this hypothesis.

One of the interesting areas of research can be the comparison of different methods and data to find the best method or data [53,54] (Rozenstein and Karnieli, 2011; Hashmi et al., 2019). Simultaneous use of data and conventional methods is also another part of the research [55,56] (Maleki et al., 2017; Alqurashi and Kumar, 2013), which are about the methods used in this research. The combination of these methods has been used for various applications [57,58,59] (Li et al., 2007; Kamali et al., 2017; Bagheri et al., 2021), as well as parts of these methods used in urban resilience research [60,61,62,63] (Banica et al., 2017; Ghajari et al., 2017; Zhang et al., 2019; Tayyab, 2021). Therefore, combining these methods to assess urban resilience can be an interesting topic. The purpose of this study is to rank the neighborhoods of Tehran in terms of resilience to determine the priority of regeneration using a combination of the Delphi and AHP methods and GIS capabilities.

2. The Study Area

Tehran is located in the southern slope of the Alborz Mountain range at 51° to 51° and 40 min east longitude and 35 degrees and 30 min to 35 degrees and 51 min north latitude [64,65]. Tehran is the capital of Iran with a population of more than ten million and is one of the largest cities in West Asia [66]. The city of Tehran consists of 22 districts and 376 neighborhoods. According to field surveys, the deteriorated urban areas are mostly in districts 10, 11, 12, and 17. On the other hand, districts 2, 5, 21, and 22 have the least deteriorated urban areas. (Figure 1) shows the study area.

In recent years, many studies have been conducted on Tehran resilience in various aspects, including Moghadas et al. [67], which focused on urban flood resilience and demonstrated that districts 6 and 22 were the most resilient and district 1 had the lowest resilience. Rezaei et al. [68] examined the resilience of Tehran from the perspective of the earthquake, with the Qeytariyeh neighborhood having the highest resilience and the Qala-e-Marghi neighborhood having the lowest. In addition, Azadeh [69] showed how improving the quality and efficiency of the rail transportation system affects the increase of resilience in Tehran. Finally, Lak et al. [70] demonstrated there is a negative correlation between neighborhood resilience and the number of coronary arteries.

3. Methodology

The research method of this study was analytical-descriptive, where the purpose was practical. The research stage included documentary and library research (further understanding of the subject, understanding of effective criteria, understanding of methods), field studies (understanding of the study area and communication with experts), and spatial (preparation of layers of spatial information and spatial processes).

3.1. Criteria Set

Preliminary criteria were determined by reviewing similar research and interviewing Iranian experts (due to the case study area in Iran). As aforementioned, resilience has many dimensions because it is used in different disciplines. However, the criteria were chosen to cover a significant range of urban resilience. Dimensions of urban resilience fall into five categories: physical, natural, economic, organizational, and social [71]. Based on these five dimensions of resilience and reviewing related research, eighteen criteria to determine the resilience of neighborhoods and areas of Tehran are presented in (

Table 1. Dimensions of Urban Resilience [71].

Dimensions	Criteria	Reference
Physical	Housing Age, type of construction materials, land use, space syntax, width of streets	[5,15,72,73,74,75]
Environmental	Green space per capita, land surface temperature, distance from fault	[5,76,77,78,79,80]
Economic	Housing prices, Income level	[5,81,82,83,84]
Organizational	Distance to the fire station, access to medical centers, density of educational spaces	[3,85,86,87,88]
Social	Population density, age composition, sex composition, residents’ education level, household density	[14,83,84,86,89]

).

3.2. Delphi Technique

Delphi is an iterative process for integrating expert consensus [90]. Delphi is a way to reach a consensus of group opinion based on the opinion of expert group members on a particular issue or problem [91]. The three main features of the Delphi technique are members’ anonymity for each other, receiving feedback, and iteration [92]. Using a single method for research in many studies reduces the efficiency of the final results and does not correspond to reality, and on the other hand, the combination of methods often improves the results. This is also true for combining multi-criteria and Delphi decision-making. The advantage of including the Delphi method with other MCDM methods is that it can initially reduce the number of indicators and decision time [93] (Wudhikarn, 2018). AHP analysis, without considering multiple methodologies (such as the simultaneous use of the Delphi approach and AHP analysis), can lead the research in the wrong direction [94] (Kim and Kumar, 2009).

In this study, the Delphi technique was used to determine the final criteria for the resilience of Tehran neighborhoods. Experts were asked to determine how important the criteria are in the discussion of resilience (in a range of seven from absolutely significant to absolutely insignificant). The criteria with the required score are then used in the final analysis (with a significance or score higher than 4). Questionnaire forms were provided to researchers who had similar research (as mentioned earlier in the criteria determination section, due to the case study in Iran, Iranian experts were used in this section). For the number of members participating in the Delphi panel in scientific research, different numbers have been used, such as 43 [95] (Iden et al., 2011), 441 [96] (Lindeman, 1975), and even 32,000. The research showed with the increase in the number of qualified experts; results are reduced [97] (Markmann et al., 2021). Due to the fact that in this study, only university experts in Iran who had similar research (making the available statistical population smaller) were used. There were 42 university experts that participated in the Delphi panel. Experts ranged in age from 25 to 65, with 24 men and 18 women. The information of the participants in the Delphi panel is presented in

Table 2. Information of the Delphi panel experts.

Expertise	Degree	Number (Persons)
Urban Design	PhD/Master	(2)/(6)
Urban Planning	PhD/Master	(3)/(5)
Crisis Management	PhD/Master	(2)/(4)
Sociology	PhD/Master	(3)/(4)
Civil Engineer	PhD/Master	(2)/(4)
Traffic Expert	PhD/Master	(1)/(2)
Natural Geography	PhD/Master	(2)/(2)
Total		42

. Furthermore, the number of repetitions and feedback to obtain homogeneous results in this study was three times.

3.3. Analytic Hierarchy Process (AHP)

After identifying the criteria, a method is needed to integrate expert evaluations in order to find the best result [98]. A set of objects that are compared in pairs must be homogeneous. That is, the priority of the most important criterion should not be more than 9 times from the least important criterion [99]. The preference in this 9-point scale is presented in (

Table 3. Pair-wise comparison scale for AHP preferences [101,106].

Preferences (Oral Assessment)	Numerical Rating
Maximum priority	9
Very high priority	7
High Priority	5
Medium priority	3
Less priority	1
Preferences between these intervals	2,4,6,8

) [100,101]. The number of AHP experts in scientific research is not specified [102,103,104] (Saaty, 1989; Saaty, 2008; Saaty a Shang, 2011). However, in a study with this statistical population, it was estimated between 3 and 200 people. This showed that if the number of experts exceeds 50, the amount of expertise will decrease, but in statistical populations (number of experts) with less than 50, experts should be specialized [105] (Tsyganok et al., 2012). Given that the experts in this study were experts who published research on resilience, 21 experts were used in the AHP department, which was different from the Delphi experts (the specialties matched, but the people were different). The age range of experts was between 25 and 65, which were 8 women and 13 men.

Duplex comparisons are performed in the form of a “pairwise comparison matrix” and that matrix is determinant by Equation (1).

$$A=\left[\begin{array}{cccc}A11& A12& \cdots & A1n\\ A21& A22& \dots & A2n\\ ⋮& ⋮& ⋮& ⋮\\ An1& \cdots & \cdots & Ann\end{array}\right]$$

(1)

In the above matrix, the amount of preferences from xi to xj is set for 0 < aij. More precisely, according to Saaty’s theory, each array of the matrix represents the approximate ratio between two weights (Equation (2)).

A_ij ≈ w_i/w_j   ∀_i˓

(2)

That is, if the arrays of the matrix represent exactly the ratio between the weights, the matrix A is in the form of Equation (3).

$$A=\left(\frac{Wi}{Wj}\right)n\ast n=\left[\begin{array}{cccc}W1/W1& W1/W2& \cdots & W1/Wn\\ W2/W1& W2/W2& \cdots & A2n\\ ⋮& ⋮& ⋮& ⋮\\ Wn/W1& Wn/W2& \cdots & Ann\end{array}\right]$$

(3)

Considering Equations (1) and (3) to the matrix, we arrive at Equation (4) so that aij = 1/aji and matrix A can be expressed in a simple and rewritten form below.

$$A=\left[\begin{array}{cccc}1& A12& \cdots & A1n\\ 1/A12& 1& \cdots & A2n\\ ⋮& ⋮& ⋮& ⋮\\ 1/A1n& 1/A2n& \cdots & 1\end{array}\right]$$

(4)

Then, calculate the CI consistency index and CR consistency ratio using Formulas (5) and (6) to confirm the effectiveness of the A comparison matrix.

$$\mathrm{CI}=\frac{ \lambda max}{n-1}$$

(5)

$$\mathrm{CR}=\frac{CI}{RI}$$

(6)

where n is the number of matrix A, RI is the mean of the random index and can be determined by

Table 4. RI rate for a number of different criteria.

N	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
RI	0	0	0.52	0.88	1.1	1.24	1.34	1.4	1.44	1.48	1.51	1.53	1.55	1.57	1.58

. If CR < 0.1, the comparison matrix meets the compatibility standard [106,107]. Comparisons are standard if the compatibility rate is 0.1 or less. The RI for a number of different criteria is presented in (Table 4).

The arithmetic mean method was used to calculate the weights [108]. The arithmetic mean is in accordance with Equation (7).

$$A=\frac{1}{n}\sum _{i=0}^n\mathrm{ai}$$

(7)

where:

A = arithmetic mean

n = number of values

ai = data set values

3.4. Data Map

Spatial analysis requires spatial information. For this reason, in order to convert the spatial data of the criteria into spatial information, relevant analysis is required. These analyzes are common in some criteria and different in others.

Table 5. Transformation of criteria into spatial information and how they affect resilience.

Criterion	Explanation	Extraction Method	Effectiveness	Ref.
Land surface temperature	In order to extract land surface temperature, Thermal Infrared Sensor (TIRS)	$T_s=T_i+C_1\left(T_i-T_j\right)+C_2{\left(T_i-T_j\right)}^2+C_0+$$(C_3+C_{4}w)$$(1-\varepsilon )$$(C_5+C_{6}w)$ Δɛ	−	[80]
Household density	Number of households per hectare	$Ho{u}_{density}=\frac{Ho{u}_{tot}}{Are{a}_{\left(ha\right)}}$	−	[109]
Population density	The number of inhabitants per hectare	$Po{p}_{density}=\frac{Po{p}_{tot}}{Are{a}_{\left(ha\right)}}$	−	[109]
Age of buildings	Average age of buildings prepared by Tehran Municipality		−
Distance to fault	Distance to fault which is calculated by using Euclidean distance	$dist\left(p,q\right)=\sqrt{{\left(x_p-x_q\right)}^2+{(y_p-y_q)}^2}$	+	[110]
Distance to fire stations	Distance fire stations that are calculated by using Euclidean distance	$dist\left(p,q\right)=\sqrt{{\left(x_p-x_q\right)}^2+{(y_p-y_q)}^2}$	−	[110]
Housing price	Determined using field surveys and price comparisons in housing finder soft wares.		−
Education center density	Kernel function estimate of density in schools	$\underset{-\infty }{\overset{\omega }{{\displaystyle \int }}}\mathrm{ḟ}\left(t\right)dt=\frac{1}{n}{\displaystyle {\displaystyle \sum }_{j=1}^x}\underset{-\infty }{\overset{\omega }{{\displaystyle \int }}}K\left(x_j,t\right)dt=1$	+	[111]
Distance to health center	Distance to health clinics which is calculated by using Euclidean distance	$dist\left(p,q\right)=\sqrt{{\left(x_p-x_q\right)}^2+{(y_p-y_q)}^2}$	−	[110]
Land use	This data was prepared by Tehran Municipality		+/−
Sex composition of population	The ratio of women to men	$sex ratio=\frac{women}{men}$	−	[112]
Street width	Distance of street center to the buildings (Euclidean distance)	$dist\left(p,q\right)=\sqrt{{\left(x_p-x_q\right)}^2+{(y_p-y_q)}^2}$	+	[110]
Space syntax	Analyzing the relationships between spaces of urban areas and buildings		+	[113]
Per capita green space	Calculation of available green space as a ratio to the number of inhabitants		+	[114]
Income level	Information was obtained by sampling and field questioning	$\mathrm{n}=\frac{t^{2}pq/d^2}{1+\frac{1}{N} \left[\mathrm{t}2\mathrm{pq}/\mathrm{d}2 - 1\right]}$	+	[115]
Level of education	This information was collected by the Statistics Organization of Iran		+
Type of building materials	This information was collected by the Statistics Organization of Iran		+/−
Age composition of population	Ratio of population under 14 and above 65 to total population	$\frac{Y+O}{Pop}$	−	[116]

summarizes the method of preparing spatial information for the criteria used. Note that in the case of the effectiveness, if it is a (+) in front of the criterion, that is, an increase in the amount of that criterion increases the productivity, but if it is a (−) sign, it means that the productivity rises by increasing the criterion. For example, the greater the distance from the fault, the greater the resilience, but the greater the distance from the fire station, the lower the resilience. Regarding land use and type of construction materials (−/+), resilience depends on the type of land use and the type of building materials.

3.5. Resilience Map and Ranking of Neighborhoods and for Regeneration

In the final step, by combining the spatial reclassed layers with the weights determined by the AHP method, a weighted map of the criteria was created. The layers were reclassified according to how they affected resilience (Table 5). The Xu method was used for reclassification, which divides the layers into five classes using mean parameters and standard deviation [117] (Xu et al., 2011). The reclassification method of this method is presented in

Table 6. Data reclassification method.

Class	Domain Related to Each Class
First class	V ≤ Vmean − 1.5std
Second class	Vmean-1.5std < V < Vmean − std
Third class	Vmean-std < V ≤ Vmean + std
Fourth class	Vmean + std < V ≤ Vmean + 1.5std
Fifth class	V > Vmean + 1.5std

. By combining all the weighted layers, a resilience map was created. The weighted linear combination (WLC) method was used to combine the layers. The WLC is the most common technique in multi-criteria evaluation analysis. In this method, the weights obtained from AHP are multiplied in the parameter map, and finally, all the weighted layers are added together [118,119,120] (Rahmani et al., 2015; Aydi et al., 2016; Zarin et al., 2021). Finally, using spatial statistics analysis, the ranking of resilience was determined, and, consequently, the regeneration operation was prioritized. Spatial statistics information is such that according to the prepared resilience map, the average value of each neighborhood in GIS is calculated; the higher the average, the more resilience, and vice versa. The lower the resilience (lower average), the greater the priority of regeneration and vice versa. Figure 2 shows the methodology of the research method used in this research.

4. Results

In this section, the results obtained from the implementation of the method on the data will be presented. In the first step, the criteria that passed the Delphi technique filter will be presented. The results of this section will determine what the focus of the research is on the processing of which spatial data, based on which the resilience ranking is determined, and, consequently, the neighborhoods to be regeneration. The final criteria for passing the Delphi technique filter to be used in the final analysis are presented in

Table 7. Final criteria for decision making.

Criteria Name
Building age(C1)	Space syntax(C6)	Green space per capita (C11)
Population density(C2)	Distance to fire station(C7)	Household density (C12)
Distance from fault(C3)	Building materials(C8)	Density of educational spaces (C13)
Age Composition(C4)	Building prices (C9)	Land use (C14)
Sex composition(C5)	Access to Medical Centers (C10)

.

According to Table 7, 14 final criteria were selected. According to the results obtained from the experts and the Delphi technique, the three criteria of land surface temperature, income level, and education level did not meet the required criteria compared to other criteria, and the criterion of street width was also excluded from the final analysis from the point of view of experts, due to the overlap with the criterion of Space Syntax. Then, 14 final criteria were entered into the AHP decision-making system.

Table 8. Pair-wise comparison matrix of criteria.

	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14
C1	1	3	2	7	8	2	3	1	6	4	9	4	5	6
C2	0.333	1	0.5	2	3	0.5	1	0.333	2	1	3	1	2	2
C3	0.5	2	1	4	4	1	2	0.5	3	2	5	2	3	3
C4	0.143	0.5	0.25	1	2	0.333	0.333	0.143	1	0.5	2	0.5	0.5	1
C5	0.125	0.33	0.25	0.5	1	0.25	0.333	0.125	1	0.5	1	0.5	0.5	1
C6	0.5	1	1	3	4	1	2	0.5	3	2	5	2	2	3
C7	0.333	3	0.5	3	3	0.5	1	0.333	2	1	3	1	2	2
C8	1	3	2	6	8	2	3	1	5	4	7	4	4	6
C9	0.1666	0.5	0.333	1	1	0.333	0.5	0.2	1	0.5	2	0.5	1	1
C10	0.25	1	0.5	2	2	0.5	1	0.25	2	1	2	1	1	2
C11	0.11	0.333	0.2	0.5	1	0.2	0.333	0.143	0.5	0.5	1	0.5	0.5	1
C12	0.25	1	0.5	2	2	0.5	1	0.25	2	1	2	1	1	2
C13	0.2	0.5	0.333	2	2	0.5	0.5	0.25	1	1	2	1	1	1
C14	0.166	0.5	0.333	1	1	0.333	0.5	0.166	1	0.5	1	0.5	1	1

is a pairwise comparison matrix of criteria that were compared according to experts.

Doing steps of the AHP method [121], the final weight of the criteria was determined, which is presented in

Table 9. The final weight of criteria.

C	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14
W	0.19533	0.06116	0.105223	0.030013	0.023893	0.096223	0.071288	0.185057	0.032192	0.052647	0.022057	0.052647	0.041953	0.030129

. According to this table, the age of the building with a weight of 19.53% is the most important criterion, and then the quality of construction materials with 18.5% is in the second place. Furthermore, according to experts, green space per capita is the least important criterion.

To investigate the validity of the results, the CR value should be calculated, which should be less than 0.1. First, the amount of CI should be calculated, which was equal to CI = 0.141795. Then, considering that 14 criteria were used, and RI is equal to 1.57 for 14 criteria, by dividing the CI by RI, we have CR = 0.090315. Given that CR is less than 0.1, the matrix is compatible.

In the next step, the criteria map was prepared according to the methods and data mentioned above, which are shown in Figure 3 and Figure 4.

The population density map (a) in six floors shows that the density is in the residential blocks, which according to the map, the population density is higher in the west and the city center. The building materials map shows that more than half of the city has good building materials (70–90% of the blocks have durable building materials). Furthermore, a large part of the city has very good building materials, with only some parts in the center of the city where less than 50% of the blocks have suitable building materials. The household density map is also very much in line with the population density map. Green space per capita shows that about 70% of the city has a green space per capita of 0 to 3 m, and a large part of the city in the center towards the outskirts of the city has a green space per capita of 0 to 1 m. In the west of the city, green space per capita reaches its highest level.

The city center to the north is often land use used for residential purposes, and the outskirts of the city are often used for mixed uses (residential-commercial, residential-industrial, etc.) or other uses such as industrial, commercial, tourism. In the city center, towards the outskirts (except the West), the sex ratio is in favor of men, but the sex ratio in the West indicates a greater number of women. In general, according to statistical data, the population of men in Tehran is more than women. The distance from the fault is shown in the six classes. According to the map, the distance from the fault is north and south of the city at a distance of less than 1000 m. The further north and south we move towards the center, the greater the distance from the fault. The population ratio map (population ratio of 0–14 people and people over 65 years to the total population) shows that most of the city level is covered by 5–10%, but the western margin is more than 20%.

Figure 4 shows the final section of the criteria map. The housing price map shows the north and south of the city show the two poles of wealth and poverty, so the term north of the city is known as a symbol of wealth and south of the city as a symbol of poverty in Tehran. The deteriorated urban areas (historic urban areas) map of Tehran shows that the deteriorated urban areas of the city are mainly in the center of Tehran and tend to the south, and there are scattered blocks in the northern part of the city. Space syntax is more in the center of the city (sloping to the west), and the closer to the outskirts of the city, the weaker the communication of the space syntax becomes. The distribution of medical spaces in the center is much higher than in other places, and the educational spaces in the west of the city are much less than in the east of the city. The situation and density of educational uses are very similar to the distribution of medical spaces, and there are more educational spaces in the center and south of the city, and east of Tehran has much more educational spaces than the west. Fire centers are also scattered in the west of Tehran, but like educational and medical spaces, it is not very crowded only in the central part, and there are many fire centers in the south, as well as in the north of the city.

In the next step, the layers were reclassified based on the importance of intrinsic values and the importance of the intrinsic value of layers in recycling. The reclassified data layers are then obtained by multiplying the weighted data layers by the AHP method by calculating the weights. Finally, the weighted layers were combined, and a flexibility map was obtained. Figure 5 shows the resilience map of Tehran based on the criteria used.

Tehran’s resilience map shows that non-resilience neighborhoods are often concentrated in three areas: (a) South of the city, (b) all over the north of the city (from northeast to the northwest), and (c) the southern part of the city center. In opposite, the resilient areas are often in two areas: (a) northern part of the city center (b) eastern zone. According to the results obtained in the resilience map, it is clear that the number of non-resilient and unstable neighborhoods is much higher than resilient neighborhoods. The proximity of unstable areas to non-resilience areas creates the problem that by making negative changes in one or more criteria, these areas also increase the sum of non-resilience areas. Looking at the resilience map of Tehran in Figure 5 and comparing it with the map of deteriorated areas, or in other words, the map of the age of the building, Figure 4B shows that there is a great deal of correspondence between these two maps. The more detailed investigation shows that the vicinity of unstable areas with these neighborhoods will make more deteriorated and unlivable urban areas in the coming years if the extensive renovation is not formed.

Table 10. Neighborhoods with the highest and lowest resilience.

Best		Worst
Rank	Neighborhoods	Rank	Neighborhoods
1	Iranshahr	357	Alaeen
2	Vali Asr Square	358	Agaahi
3	University of Tehran	359	Taqiyabad
4	Ghaem Magham-Sanai	360	Shoosh
5	Ismailabad	361	Rah-Ahan
6	Laleh Park	362	Saffaeh
7	Amjadieh-Khaghani	363	Abuzar
8	Ferdowsi	364	Southern Armenians
9	Imamate	365	Moghadam
10	Parastar	366	Anbar Naft
11	Choobtarash	367	Firoozabad
12	Behjat Abad	368	Golchin
13	Zanjan	369	Imamzadeh Yahya
14	Argentina	370	Mansoorieh
15	Eastern Tehranpars	371	Zahirabad
16	Western Tehranpars	372	Shahid Ghayouri
17	Tehran Pars	373	Dilman
18	Palestine	374	Jalili
19	Jahad square	375	Valiabad
20	Shoora	376	Sartakht

demonstrates the ranking of 20 neighborhoods with the highest resilience and 20 neighborhoods with the lowest resilience. According to this table, the neighborhoods of Iranshahr, Valiasr Square, the University of Tehran, Ghaem Magham-Sanai, and Ismailabad are the five neighborhoods with the highest resilience. On the other hand, Sartakht, Valiabad, Jalili, Dilman, and Shahid Ghayuri neighborhoods were the five neighborhoods with the lowest resilience, respectively. This table shows the 20 neighborhoods with the lowest resilience; this includes priorities for the regeneration process (from 357 to 376).

The point to be noted is that the status of criteria in neighborhoods that have low resilience and need to be rehabilitated should be determined and compared with the general situation of the city to determine the roadmap for regeneration and increased resilience. (

Table 11. Comparison of non-resilient neighborhoods with the whole city.

Sub-Criteria	City Situation	Status of Non-Tolerant Neighborhoods	Sub-Criteria	City Situation	Status of Non-Tolerant Neighborhoods
C1	14.6%	78.4%	C8	85%	68%
C2	249	291	C9	9.96	18.67
C3	6250	4050	C10	1250	970
C4	0.01	0.128	C11	16.5	3.5
C5	85.8	85.94	C12	75	84
C6	0.428	0.3	C13	1.69	2.31
C7	4240	3460	C14	39%	51%

) shows the status of the criteria used throughout the city and the 20 neighborhoods with the lowest resilience.

Table 11 shows that in terms of deteriorated urban areas (C1), there is a huge difference between the whole city and 20 neighborhoods in need of regeneration. More than 78 percent of non-resilience neighborhoods are deteriorated urban areas, compared with less than 15 percent for the city as a whole. In terms of population density, there is not much difference between the city (249) and the selected areas (291). Regarding the average distance from the fault, the average distance from the fault in the whole city is 2.2 km less than the selected areas. The parameter that has the most difference is the age ratio. The sensitive population (less than 15 years and more than 65 years) in selected areas is about 13 times. Regarding the other parameters, the space syntax in the city is 1.5 times better than the areas in need of regeneration. The building materials reported were 85% of the whole city, and 68% of the selected areas had good quality materials. The average property price on a scale of 22 (number of areas of Tehran) is about 10 for the whole city and 18.67 for selected areas (1 was the highest price and 22 was the lowest price). The per capita green space in the whole city is about five times that of the selected areas, and for the land use criterion, 39% of the whole city has residential use (residential only), which is 51% for the selected areas. The selected areas were in a better position than the city in the two parameters of sex ratio and distance to medical centers. For the remaining criteria, there is not much difference between the two areas, although the situation of 20 selected neighborhoods in these criteria is also worse than the average of the whole city.

5. Conclusions

Nowadays, cities are known as the largest place of human community and also the most attractive place to attract a population. For this reason, cities and their living conditions have attracted the attention of many researchers. These studies include: urban quality of life [122,123], urban growth [124,125], urban crime [126,127], urban economy [128,129], and urban climate [130,131]. If we consider all the criteria together, we will reach resilience, sustainability, and urban regeneration.

The aim of this study was to investigate the resilience of Tehran and consequently prioritize neighborhoods in terms of sustainability and regeneration. The Delphi technique helped identify and screen the most effective criteria. This technique helps to avoid reviewing and processing criteria that have a significant impact on the goal, which saves time and money spent on projects and research. The use of the Delphi technique caused the four criteria of income level, education level, street width, and land surface temperature, which had lower priority compared to other parameters, not to be considered in the final analysis.

The results of the AHP analysis on the data showed that the most important criterion in the resilience of Tehran was first the building age and then the building materials. This indicates that in the discussion of regeneration of non-resilience neighborhoods, the historic urban areas should be regenerated first, and during the regeneration process, high-quality construction materials should be utilized so that the neighborhoods achieve an acceptable level of resilience and sustainability. On the other hand, green space per capita, sex composition, and age composition had the least impact. One solution to increase per capita green space in areas where extensive change is not possible in the city can be the use of green roofs [132]. However, even if the other two factors had high effects, it was very difficult to make spatial changes in these two factors. Among the 14 criteria used, the distance from the fault does not change at all, and then the sex composition and age composition, which is very difficult to make spatial changes in these parameters. Regarding the other criteria, the changes are more or less applicable.

The final results revealed that if we divide the city into two parts, north and south, all 20 neighborhoods with the lowest resilience were located in the south of the city. This is evident in the beliefs and even the discourse of the citizens, who consider the south of the city as a symbol of poverty, low quality of life, lack of urban health, and low citizenship rights, and the opposite is true for the north of the city [133]. More interestingly, all 20 neighborhoods had the highest priority for regeneration in the deteriorated areas of the city, which shows the importance of regeneration in the historic urban areas.

The final analysis of the criteria in Table 9 showed that 78% of the buildings in the 20 non-resilience neighborhoods had deteriorated urban areas while being only 14% for the city buildings (of which 20 selected neighborhoods are part). The criterion that had the most difference was the age ratio, where the sensitive age group in non-resilience neighborhoods was about 13 times the whole city. This often stems from the birth rate in slums and their turning into Child labor. The status of the criteria is determined according to Table 9. However, the point to be noted, given the large difference in property prices in the north and south of the city, is that the government can adjust this criterion by enacting laws and structures, and by adjusting this criterion, we will see a change in some other criteria, such as age and sex ratio.

Finally, some suggestions can be made as a roadmap for future research. One of these suggestions is to re-measure the resilience status by changing the number of criteria, in other words, increasing or decreasing the criteria. Another suggestion for research is that by improving the criteria of non-resilience neighborhoods and performing re-analysis, changes in their degree of resilience can be measured. The results of this research can help managers to know how much change in what parameters can create a sustainable city. Furthermore, by using other multi-criteria decision-making methods and matching its results with this research, higher levels of reliability can be achieved.