A Comprehensive Study of the Suitability of Urban Underground Spaces for Connection Development: A Case Study of the Erhai Lake Basin, China

Abstract

Urban underground space (UUS) involves multiple stakeholders whose concerns span the entire life cycle of underground space. The government pays more attention to the social benefits of UUS to the region, developers pay more attention to the economic benefits brought by the development of UUS, and users pay more attention to the comfort level of UUS operation. This study used the analytic Hierarchy Process (AHP), government, developers, and users to construct a comprehensive evaluation index system of underground space development needs. Different from previous studies, considering the need for future connectivity development in the UUS circle, this paper proposes to comprehensively consider the impact of connectivity development on regional suitability through coupled AHP and cellular automata (CA). The results show that about 102.67 km² (13.03%) underground area of Erhai Lake Basin is suitable for underground space development. Compared with the traditional evaluation method, the evaluation result of about 31.41 km² (3.99%) area was improved, and the fragmentation problem between blocks was improved. This method is of great value to the comprehensive development of UUS in the Erhai Lake Basin and is helping to improve future underground space planning.

1. Introduction

By the end of 2020, China’s urbanization rate of permanent residents had exceeded 60%. Rapid urbanization brings many social benefits but also many problems, such as space resource shortage, traffic congestion, and environmental pollution [1,2]. The development of urban underground urban space (UUS) increases space resources and can help alleviate these problems. Therefore, the development of UUS is considered a viable solution to shortages of urban spatial resources [3,4,5,6], leading to UUS receiving increasing attention [7]. However, to date, the development of UUS has been haphazard. Historically, development was mostly based on a first-come, first-served strategy, which resulted in severe damage to UUS resources [8,9]. To solve this problem, it is very important to study the utilization of UUS and discover and understand its underlying mechanism.

The development of urban UUS is a complicated process, which is affected by many aspects such as social and economic benefits [10,11,12,13,14,15,16]. Due to the types of evaluation subjects, most of the existing evaluation studies focus on the impact of a certain aspect, thus ignoring the feelings of other stakeholders [17,18,19,20,21,22,23]. To overcome this problem, the efficient development and utilization of the city can be promoted by integrating the concerns of various stakeholders [24]. AHP can well solve the problem of multi-stakeholder index construction [25]. In 2023, Adhi and Muslim analyzed and studied fuzzy AHP for owners, contractors, consultants, subcontractors, architects, government, local government and NGOs, and other stakeholders of the gap between the analysis and research. Finally, the sustainable building implementation plan was developed [26]. In 2023, Aldossary et al. analyzed and evaluated the sustainable urban development projects in the Baja region by combining the built environment, engineering, and investment field decision-makers [27].

UUS does not exist solely in isolation but is often interconnected, and the internal economy that this creates is also another factor driving demand for the development of future UUS. The degree of connectivity between different neighboring UUS regions will affect their vitality. Therefore, the suitability of UUS development is not only affected by the economic conditions in the region in which the UUS is located but also by the economic conditions of neighboring areas. The existence of adjacent neighborhoods suitable for UUS development will promote the vitality of the city center, enhancing its suitability for UUS development. To consider the influence of interconnectivity, a suitable model is needed for simulation research. The Cellular Automata (CA) approach is well-adapted to this influence relationship, and CA has been widely used in urban research [28,29,30,31]. In 2018, Liang, X. et al. coupled system dynamics (SD) and CA models in the Future Land Use Software (FLUS) model during the projection period. Taking the Pearl River Delta region of Southern China as an example, they explained how the combined approach generated urban patterns under different scenarios [32]. In 2021, Zhang, Y. et al. established an intelligent planning model combined with CA and evaluated the development and utilization of UUS in the Luohu District of Shenzhen, Guangdong Province, China, by analyzing the geological conditions and the current situation of surface construction [33]. It is believed that the connection development of neighboring UUS can be considered by integrating the excellent spatiotemporal simulation and evolution ability of CA to help make a more appropriate evaluation of UUS development.

Previous UUS-related studies only focused on index construction from the perspective of some groups, ignoring the interest correlation between other stakeholders and UUS [34,35]. Therefore, it is necessary to build a Multiple perspective UUS evaluation index system from the perspective of UUS. By analyzing the UUS indicators concerned by various stakeholders, the corresponding index system is constructed, to increase the participation of stakeholders in the UUS development program. The connectivity among UUS is also a good way to solve the fragmented distribution of UUS and help the development of UUS with high spatial vitality. Improving the index construction and connectivity analysis from the perspective can help the sustainable and healthy development of urban space and bring more development power to the development of cities.

To holistically address the needs of numerous stakeholders and to also account for the economic benefits emerging from linking adjoining UUS, this study advocates for the establishment of a comprehensive stakeholder index system for the Erhai Lake Basin. The said system will facilitate the appraisal of the value of UUS within the region. The evaluation classes are fit to a fine grid to allow a convenient and more accurate quantitative analysis of the status of the study area. The grid construction also facilitates the simulation and analysis of UUS in the later stage. Integrating the preferences of various stakeholders makes the solution more relevant to the demands of society. In contrast, the use of CA simulation analysis makes the final results more in line with the future development of the Erhai Lake Basin. This work aims to is to generate suggestions for the development of UUS in the Erhai Lake Basin and demonstrate the usefulness of this approach to the comprehensive development of UUS in other urban centers.

2. Materials and Methods

2.1. Overview of the Study Area

Located in Dali Bai Autonomous Prefecture in the northwest of Yunnan Province and at the southern end of the Hengduan Mountains, the Erhai Lake Basin of Dali Prefecture is a plateau basin bound by mountains and rivers. The Erhai Lake Basin includes 18 towns and townships in Dali City and the surrounding Eryuan County. The Haixi area includes communities such as Dali and Xizhou Towns, and the southern metropolitan area includes communities such as Xiaguan and Fengyi towns. In contrast, the mountainous coastal area includes communities such as Shuanglang and Wase Towns. Finally, the upstream area of Erhai Lake includes communities such as Cibihu and Yousuo Towns. The actual area is 1087 square kilometers, of which 70% is mountains, 15% is water, and 15% is the basin itself (Figure 1). Given that the entire study area surrounds the Erhai Nature Reserve, which is intended for the regeneration of the ecological and scenic environment, several construction and human activities, such as the creation and extension of residential regions and the admission of non-tourist motor vehicles, are limited. At the same time, there is a need for tourist facilities and access. As the development of UUS in the Erhai Lake Reserve is similarly restricted, the Erhai Lake Reserve is not included in the study area.

This study can help the development of the Erhai Lake Basin in other ways. Identifying high-quality UUS development zones can help establish new production and living spaces, accelerate the creation of a sustainable society, promote the rational utilization of land and space, and play a positive role in the effective protection of the environment. Furthermore, resolving the developmental obstacles of the Erhai watershed while being mindful of available resources and ecological considerations can promote sustainable development in other regions that necessitate UUS development to mitigate the stress on urban land usage. The specific process of evaluating the suitability of UUS development in the Erhai Lake Basin adopted in this study is shown in Figure 2.

2.2. Index Selection

When selecting indicators, we collected more than 40 articles related to existing research on the development suitability of UUS in China, and studies that considered regions in most provinces of China (Figure 3). Drawing upon the peculiarities of the research area, we ultimately chose three criteria—spatial advantage, commercial gain, and comfort level—for examination. As these three indicators are the main concerns of the government, developers, and users, respectively, the index system can comprehensively consider the needs of multiple stakeholders.

The indicators were selected based on the study of the relevant literature combined with the opinions of experts and local residents who are familiar with the situation in the study area. The spatial advantage metrics primarily entailed population density, traffic status, and building density. On the other hand, commercial benefit indicators comprised land utilization patterns, commercial land value, and location. The comfort level indicators involved road traffic accessibility, bus stop density, and the extent of point-of-interest (POI) blending (as illustrated in Figure 4). These indicators are described in detail below.

Spatial benefit: the space resources brought by the development of UUS can effectively help alleviate traffic congestion, a shortage of space resources, and other problems and help promote sustainable development. Therefore, space benefits brought by the development of UUS are one of the main issues that the government is concerned about [36]. Indicators such as population density, traffic conditions, and building density are closely related to how efficiently space resources are used.

Population density C1: population density is the main indicator reflecting the vitality of a region. Areas with high population density are dynamic and serve as the primary hubs for urban advancement. However, they are also vulnerable to spatial resource constraints. UUS development can offer spatial benefits and aid in mitigating the burden on spatial resources arising from high population density.

Traffic conditions C2: traffic conditions are a key concern in modern urban development Traffic is the main artery of urban vitality, and congested traffic hinders the efficient and orderly development of a city. To relieve the pressure caused by traffic congestion, in addition to building overpasses, asking for space underground for use as underground tunnels is an effective approach.

Building density C3: areas with high building density occupy land that could otherwise be used for agriculture and transportation, which deteriorates the regional environment and is not conducive to the future sustainable development of the region. Through UUS development, certain functions currently conducted by above-ground constructions can be shifted below ground, reducing various challenges resulting from a superfluous concentration of above-ground buildings.

Commercial benefit: the development of UUS also has great commercial value, which is one of the primary motivations of property developers. Centering the development of UUS in areas with high commercial value brings higher commercial benefits. The most important factors that influence commercial benefits are land use types, commercial land price, and location [37].

Land use types C4: different land use types have different characteristics. Commercial land and residential land contain great passenger and human flow. The development of UUS in these areas can both add commercial value and improve the vitality of the region to help the future development of the region. The development of UUS for land types with great commercial value is one of the primary concerns of developers.

Commercial land price C5: areas with high commercial land prices are bound to have a superior geographical location, a more pleasing environment, and better development potential. UUS development can also provide additional spatial resources to facilitate developers in attaining substantial profits. Therefore, UUS development around high commercial land price areas is another major concern of developers.

Spatial location C6: in the process of city development, a clear development center will be established, and future developments will be based around this center, which is generally divided into the main development center and some sub-level development centers. Development around the city center is also an inevitable demand in urban development and something that developers need to focus on.

Comfort level: for user groups, the most evident concern during the usage of UUS is their comfort. UUS with a high comfort level can make users feel happy, while UUS with a low comfort level will make users feel depressed and affect their later experiences [38]. Therefore, UUS comfort is the main concern of users.

Road traffic accessibility C7: road traffic is the main artery of the city, adding both value and transportation convenience. Areas with good accessibility are convenient for users to travel to and around and can deliver a more comfortable user experience. Therefore, road traffic accessibility is one of the other major concerns of users.

Bus stop density C8: public transportation in the Erhai Lake Basin is dominated by bus travel, and public transportation is the main choice for daily travel for most residents. An area with a high density of bus stations offers greater travel convenience and enhances user comfort to a certain degree. Correspondingly, bus stop density is a key element in determining UUS comfort.

Degree of POI mixing C9: the degree of mixing of POI is often an important decision factor when users go to regions for consumption. Areas with a high degree of POI mixing deliver a better experience to users, and various types of user groups can be attracted, thus, enhancing the comfort level of the UUS.

The data used in this paper came from sources, including the Overall Urban Planning of Dali City Annual Reports (2011–2020) and the Dali Statistical Yearbook (2010–2020). The original data map was drawn using data from the map of China (GS (2019) 1822), the 2020 vector map of the population density of the Erhai Lake Basin, and POI data of the Erhai Lake Basin. The basic geographic information data of the 1:1 million public version (2021) and the 1:250 thousand National Basic Geographic Database (2015) were utilized. The data and data types used to establish the indicator system are shown in

Table 1. Research data sources and data typology.

Evaluation Data	Source	Data Typology
Population density (C1)	Population density of Erhai lake basin City— www.worldpop.org (accessed on 10 June 2022); (Resolution: 1 km × 1 km); (Year 2020)	Raster
Traffic conditions (C2)	Baidu Map transportation big data platform—www.jiaotong.baidu.com (accessed on 10 June 2022); (Linear data); (Year 2020)	Raster
Building density (C3)	1:1 million public version of basic geographic information data (Year 2021), 1:250,000 National Basic geographic database (Year 2015)	Vector
Land use types (C4)	The third national land resource survey provided by Kunming City Planning and Information Center (Year 2019); (Resolution: 50 m × 50 m)	Raster
The commercial land price (C5)	Housing transaction data from 58 cities and HOME LINK net (Year 2020); (Resolution: 200 m × 200 m)	Raster
Spatial location (C6)	Dali City Master Plan (Year 2011–2020); (Resolution: 100 m × 100 m)	Vector
Road traffic accessibility (C7)	National Basic Geographic database of National Catalogue Service for Geographic Information—www.webmap.cn (accessed on 10 June 2022); (Year 2020); (Linear data)	Vector
Bus stop density (C8)	Baidu Map and Amap traffic station POI data (Dotted data); (Year 2020)	Raster
POI mixing degree (C9)	POI data of Dali (Dotted data); (Year 2020)	Vector

, and a sample of the original data is shown in Figure 5.

After the data had been collected, to facilitate more accurate quantitative analysis to reflect the value of the UUS development, the evaluation model of UUS development and suitability evaluation standard divided Dali City into four levels, with level 1 representing space inappropriate for development, level 2 space less suitable for development, level 3 space more suitable for development, and level 4 space most optimum for development. Based on the Master Urban Planning of Dali City (2011–2020), the study area’s specific conditions, expert opinions, and pertinent rating standards, the evaluation standards for the selected metrics were formulated (

Table 2. Index evaluation standard.

Index Type	Quantitative Method	Index Classification
		Level 1	Level 2	Level 3	Level 4
Population density C1	X1 = NC1/SC1 (km2) NC1 is the population; SC1 is the area of the region.	X1 ≤ 1000	1000 < X1 ≤ 1500	1500 < X1 ≤ 2000	2000 < X1
Traffic conditions C2	X2 = tC/tF tC is the time taken during the congestion period; tF is the time spent during the unblocked period.	1 < X2 ≤ 1.5	1.5 < X2 ≤ 1.8	1.8 < X2 ≤ 2.0	2.0 < X2
Building density C3	X3 = mB/mS mB is the total basal area of the building; mS is the area of planned construction land.	X3 ≤ 0.1	0.1 < X3 ≤ 0.25	0.25 < X3 ≤ 0.4	0.4 < X3
Land use types C4	It is classified according to different land use types.	Wetland, protective land, water areas, etc.	Land for residence, education, medical treatment, etc.	Logistics, warehousing, industrial land, etc.	Commercial offices, shopping mall plazas, etc.
The commercial land price C5	Based on the average price of land transactions in the region (¥/m2)	X5 ≤ 2000	2000 < X5 ≤ 3000	3000 < X5 ≤ 4000	4000 < X5
Spatial location C6	Based on the scope of the planned area in the relevant plan.	Non-urban central impact area.	Not the center of the city, but affected by the center.	Regional city center.	Center for Overall Urban Development.
Road traffic accessibility C7	Based on which type of road radiation the area is in, it is obtained through the buffer.	All kinds of roads do not radiate into this area.	In the provincial roads, national roads, township roads, and other road radiation areas.	In the radiation zone of class I and class II roads.	In the radiation zone of class III and class IV roads.
Bus stop density C8	X8 = nC8/SC8 nC8 is the number of regional bus stops; SC8 is the area of the grid (2.25 hm2).	X8 ≤ 5	5 < X8 ≤ 1 0	10 < X8 ≤ 15	15 < X8
POI mixing degree C9	$X_9=-{\displaystyle \sum _{i=1}^n}\left(P_i\log P_i\right)$ Pi is the probability that POI is i; The higher the X9 in the region, the higher the POI mixing degree.	X9 ≤ 0.2	0.2 < X9 ≤ 0.4	0.4 < X9 ≤ 0.6	0.6 < X9

). Subsequently, utilizing the established index evaluation standards, the data were categorized, sorted, and the corresponding attribute table was created in ArcGIS to perform data classification computations. Thereafter, individual index classification maps were generated for each metric (Figure 6).

2.3. Research Methods and Models

This study comprehensively evaluated the suitability of different UUS in the Erhai Lake Basin for connection development by combining AHP and CA. To establish a hierarchical structure model, the AHP was employed to combine the perspectives of stakeholders and subsequently evaluate the significance score of each metric to assign weightage to each influencing factor. The evaluation results based on AHP were then obtained by stacking analysis of the corresponding weights of each index layer through ArcGIS. However, at this stage of the analysis, the rating results did not take the tandem effect of connected underground spaces into account, and opportunities for connection development impact the final evaluation results. After scrutinizing and examining the gathered data of the respective developed underground spaces, the optimal cell configuration was chosen, and parameters were established in Matlab. Subsequently, the impact of neighboring cells was taken into account for cell simulation analysis to derive the ultimate comprehensive evaluation outcome concerning the suitability of connecting and developing adjacent underground spaces in the Erhai Lake Basin. The specific steps of AHP and CA were as follows:

2.3.1. Steps of the AHP

Build a Hierarchical Structure Model

The AHP needed to divide the decision-making objectives, factors considered (decision-making criteria), and decision-making objects into the highest, middle, and lowest levels according to their mutual relations. Through consideration of the relative weight of the lowest layer to the highest layer, various schemes and measures in the lowest layer were sorted according to the relative weight to make choices or form choices among different schemes. Thus, constructing the hierarchical structure model constituted the initial phase. Drawing upon pertinent research data regarding the evaluation and fuzzy comprehensive evaluation of UUS development suitability assessment in China, this study integrated the three criterion layers of the assessment system, namely, spatial benefit, commercial benefit, and comfort level. According to the specific requirements of the criterion layer, the criterion layer was refined, and nine specific evaluation indexes were selected as the basis for determining the degree of suitability of UUS development and establishing the evaluation index system of UUS development value in the Erhai Lake Basin (

Table 3. Evaluation index system of underground space development value.

The Target Layer	Criterion Layer	Index Layer
Suitability of underground space development (A1)	Space benefit (B1)	Population density (C1)
		Traffic conditions (C2)
		Building density (C3)
	Commercial benefit (B2)	Land use types (C4)
		The commercial land price (C5)
		Spatial location (C6)
	Comfort level (B3)	Road traffic accessibility (C7)
		Bus stop density (C8)
		POI mixing degree (C9)

).

Construct the Comparison Matrix

The hierarchical structure model is a comparison matrix. The construction of a comparison matrix requires the rater to be familiar with the importance grade of the comparison between indicators, and it is based on a judgment matrix. In this process, it is advisable to use relative scales to minimize the complexity of comparing various factors and enhance precision. The pairwise comparison Saaty scaling method widely used in the AHP is shown in

Table 4. Compare matrix scoring criteria.

Definition	Instructions	ai/aj
Equally important	i is as important as j	1
Low degree of important	i is a lower degree of importance than j	3
Medium degree important	i is a medium degree of importance relative to j	5
Highly important	i is highly important relative to than j	7
Extremely important	i is extremely important relative to j	9

:

Questionnaires were distributed according to the scoring standards, and, after the index scores were collected, the following comparison matrix could be constructed:

$$A={\left[a_{ij}\right]}_{n\times n}=\left[\begin{array}{ccc}{a}_{11}& \cdots & a_{1j}\\ ⋮& \ddots & ⋮\\ a_{i1}& \cdots & a_{ij}\end{array}\right]$$

(1)

where n is the matrix dimension, i, j = 1, 2... n.

Consistency Check

In the actual implementation of AHP, to ascertain the scientific validity of the matrix, it is essential to verify the consistency of the judgment matrix. Based on the maximum characteristic root λ_max of the judgment matrix, the consistency index CI and the consistency ratio CR are introduced through the following:

The consistency index CI is calculated from

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

(2)

When CI = 0, matrix A is as in Formula (1), while the larger CI is, the more serious the degree of inconsistency of matrix A.

The consistency ratio is calculated from

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

(3)

The consistency of the judgment matrix is considered acceptable when CR < 0.1, and the values of RI can be found in

Table 5. Random consistency index RI.

Matrix Order (n)	1	2	3	4	5	6	7	……
RI	0	0	0.58	0.90	1.12	1.24	1.32	……

.

Determination of Weight

After the consistency of the matrix had been checked, the weights could be analyzed. According to the characteristics of the largest eigenroot λ_max of the matrix, it can be known that:

$$AW={\lambda }_{max}W$$

(4)

where W is the final ranking weight vector, and each element in the vector corresponds to the weight proportion of each factor.

$$w_{1,2}=w_1\times w_2$$

(5)

where w_1,2 is the final weight of the indicator layer, w₁ is the weight of each criterion layer, and w₂ is the weight of the indicator layer in the current criterion layer.

2.3.2. Steps of the CA

CA are dynamic models with discrete distributions of time, space, and state, calculated through spatial interaction and temporal causality. They can simulate the spatio-temporal evolution of complex systems [39,40,41]. By factoring in the spatial configuration of the locality, the parameters established for AHP were rectified. To achieve this, the CA model was utilized to replicate the impact of interconnected neighboring UUS and facilitate the production of the ultimate evaluation outcome after adjusting the AHP evaluation result.

Cell Construction

Before cell analysis, the cell must be constructed, a process that can be divided into two parts: selection of cellular neighborhood pattern and selection of cellular size.

Selection of cellular neighborhood pattern: there are two types of neighborhood spatial structure commonly used in CA, Von Neumann neighborhoods and Moore neighborhoods, with the standard Moore neighborhood being the most used in urban research (Figure 7). The standard Moore neighborhood refers to the neighborhood range comprising cells within the adjacent square area. Considering the characteristics of urban development, the influence of the surrounding area is limited, so the standard Moore neighborhood was used in this study to construct the regional cell.

Selection of cellular size: Based on data for the extent of underground commerce and parking lot development in Kunming, the capital city of Yunnan Province, the average development area of Kunming is 2 hm², and the future development of the Erhai Lake Basin is likely to follow the same pattern. Based on this analysis, the size of the cells in the Erhai Lake Basin was selected as 150 m × 150 m, and the final area of the whole study area was divided into a 592 × 296 grid, in which a total of 35,034 cells had attributes.

Construction of CA Rules

Following the establishment of the cell, the corresponding CA rules were defined. The operational regulations of the cellular model adopted in this study are outlined below:

When considering the connection development of neighboring UUS, the suitability of cell X is improved by the efficiency value:

$$S\left(\mathrm{x}\right)=P_{neigh}\left(t\right)\times P_{type}\left(t\right)\times w_a$$

(6)

where P_neigh(t) is the influence value of the suitability of UUS in the neighborhood, P_type(t) is the influence value of the evaluation of the type of UUS suitable for development in the neighborhood according to the main type of land use in the region, and w_a is the distance weighted score of the neighborhood to the center.

The change of cellular state depends on the value of S. When the value of S surpasses the predetermined RA coefficient value, the corresponding cellular state transforms. Conversely, if the value of S is lower than the specified RA value, the relevant cellular state remains unaltered. The corresponding cellular states are analyzed as follows:

$$S=\left\{S1,S2\right\}$$

(7)

$$S=\{\begin{array}{c}S1, S\left(X\right)\ge RA\\ \\ S2, S\left(X\right)<RA\end{array}$$

(8)

where S1 is a constant state, S2 indicates that the suitability is improved, and RA is the reference value obtained from sample training.

After cell construction, the rules of the cell were determined, that is, the value of RA was determined. Based on the statistics of the collected information of all existing UUS developments, the operation rules of the cell were finally determined considering the different interconnectivity types of the UUS. There are two possible scenarios of cell structure (Figure 8): In scenario A, the development suitability score of UUS in the surrounding connected area is high, the land type score is also high, the comprehensive evaluation is higher than RA, and the suitability of the central cell is improved. In scenario B, the development suitability score of UUS in the surrounding connected area is low, the land type score is poor, the comprehensive evaluation is lower than RA, and the suitability of the central cell is not improved.

3. Results

By performing the computation and analysis of the scoring matrix, the weightage of each criterion layer and index layer was derived (

Table 6. Evaluation results of the weight of each index.

The Target Layer	Criterion Layer	w1	Index Layer	w2	w1,2	Order
Suitability of UUS (A1)	Space benefit (B1)	0.3149	Population density (C1)	0.3180	0.10013820	6
			Traffic conditions (C2)	0.2910	0.0916359	7
			Building density (C3)	0.3910	0.1231259	4
	Commercial benefit (B2)	0.4096	Land use types (C4)	0.2667	0.10924032	5
			The commercial land price (C5)	0.3333	0.13651968	3
			Spatial location (C6)	0.4000	0.16384000	1
	Comfort level (B3)	0.2755	Road traffic accessibility (C7)	0.2925	0.08058375	8
			Bus stop density (C8)	0.2038	0.05614690	9
			POI mixing degree (C9)	0.5037	0.13876935	2

). As illustrated in the table, indicators such as spatial location and the degree of POI mixing carry significant weightage, whereas the density of bus stops and accessibility of road traffic are relatively less significant. Following the acquisition of the weights, ArcGIS was employed to allocate weightage to each indicator layer, and the evaluation of the three criterion layers, i.e., spatial benefit, commercial benefit, and comfort level, could be superimposed. The evaluation result of the AHP could then be obtained by superimposing the target layer again (Figure 9). At this stage of the analysis, the result shows that UUS of about 71.26 km² (9.04%) is suitable for development. From the perspective of spatial distribution, the regions suitable for UUS development are mainly distributed in the south of the region, which is the core region and the main center of regional politics and economy. The performance of the remaining regions is poor; most of these regions are less developed areas, and regional population density and commercial value are low. Nonetheless, this assessment outcome lacks the analysis of the interconnectedness of underground space, and the forthcoming UUS is expected to be a collectively developed subterranean business district that requires consideration of the impact of the surrounding neighborhood on the central area. Therefore, the evaluation result must be revised via CA.

We established comparative analysis samples according to AHP analysis results, analyzed AHP samples and CA results, and modified CA transfer rules. Based on the constructed cell type and cell rules, the cell matrix was input into MATLAB, and the cell rules were input for cell simulation correction. The revised evaluation result is shown in Figure 10. Based on the revised outcomes analysis, it was revealed that out of the 31,273 cells, representing an area of 703.64 km² (89.27%), none of them demonstrated any changes in the suitability of UUS development. On the other hand, a total of 3760 cells, corresponding to an area of 88.1 km² (10.73%), exhibited an enhancement in the suitability of UUS development. The revised results show that an area of about 102.67 km² (13.03%) of the Erhai Lake Basin is suitable for UUS development. As the result shows, the suitability of UUS development in the cells of each regional development center has been improved due to the inclusion of the influence of the surrounding neighborhood, and these regions have the potential to become key areas for future development. Planning UUS as a whole is the future development trend, and to avoid improper utilization of UUS resources, the role of connected development must be considered.

Previous studies have focused on the study area and ignored the influence of neighboring, potentially interconnective, UUS. The use of CA simulation can help us fully consider the influence of the surrounding neighborhood, resulting in an improvement in the suitability of many areas. Taking into account the impact of the surrounding neighborhood, the expanded area suitable for UUS development can become an integral part of the underground economy. Therefore, it is crucial to incorporate the concept of interconnected UUS into the planning of future UUS development. By combining relevant development laws, long-term planning applicable after the end of construction can be formulated, and the evaluation map of future development obtained, thereby obtaining a result more in line with the actual development of the city.

4. Discussion

4.1. Discussion on the Status Quo of UUS Development

UUS plays an increasingly important role in urban development. As an important part of urban space, the scientific development of UUS has become the goal of both developed and developing countries [42,43,44]. In the development of UUS, the previous evaluation mostly took the quality of resources, capacity, and development potential as the evaluation factors. Zhu, H.H. et al. (2023) used digital technology to evaluate the location of UUS in Changzhou by taking engineering elements as indicators. The author mainly analyzes and studies the difficulty of engineering implementation and pays attention to the cost of development [45]. Zhang, Y.B. et al. (2023) analyzed and studied the suitability of UUS development in Kunming City by combining the social and economic value of UUS development. From the perspective of the government, the author considers how to promote urban development through the rational development of UUS [46]. With the deepening of UUS research, it becomes extremely important to consider the perspectives of different stakeholders in urban development.

In addition to the index system, the traditional UUS evaluation units are mostly single blocks. With the continuous deepening of urbanization, the interaction between blocks has also attracted wide attention. Sharifi, A. introduced the development of future cities from the aspect of neighborhood units and emphasized the importance of neighborhood units [47]. Mehaffy, M. et al. studied urban neighborhood units to help cities develop sustainably in the future [48]. The high vitality and healthy development of future cities cannot be separated from the analysis and discussion of neighborhood units. For the strategic layout of UUS, the development of urban space is needed, and the neighborhood analysis of UUS is also very necessary.

To conduct a more comprehensive analysis of the suitability of UUS development, it is necessary to consider these problems. Based on the characteristics of the abovementioned studies, this paper constructed an index system from the perspective of different stakeholders and considered the influence between blocks to obtain the final evaluation results of the Erhai Basin, which can help the sustainable development of urban UUS resources.

4.2. The Significance of Multistakeholder Indicator System

UUS is a complex urban complex, involving underground transportation, underground commerce, and underground municipal services. The life cycle of the UUS includes the development of an underground space policy, the implementation of underground engineering, and the operational management of the facility. The government sells the right to use the land according to the specific use of the city plan. To derive profits, the developer performs targeted development in the region. In the operation of UUS, users have an intuitive feeling of the layout of UUS. The three parties have different concerns throughout the UUS lifecycle.

By combining the perspectives of the government, developers, and users, this study represents three stages of underground space policy formulation, project implementation, and operation respectively. The three perspectives focus on the social benefits, economic benefits, and comfort brought by the development of UUS. In the real world, the combined influence of these factors constitutes a complete perspective of evaluating underground space. Multi-perspective evaluation of UUS is beneficial to the rational development and utilization of UUS.

4.3. Consider the Implications of UUS Connectivity

The dispersion of the UUS layout is a major problem troubling the development of UUS. In recent years, the interconnected projects of UUS have been promoted successively, which proves that connected underground space has a lot of social and economic benefits. In the past, the connectivity degree between UUS was low, mainly because the benefits brought by connectivity were not taken into account in the planning and design stage of underground space, which led to the difficulty of implementing connectivity in the later stage and the difficulty of engineering. The underground space with low connectivity is bound to restrict urban development. To meet the transformation of UUS from quantity growth to quality optimization, the underground space planning model considering connected development will become the future trend.

In the past, the traditional UUS evaluation method caused the separation of the evaluation blocks and ignored the mutual influence of the blocks. Considering the importance of connection development, this study simulates and modifies the preliminary evaluation results of UUS by combining the CA model. Compared with the traditional UUS evaluation model, the coupled model obtains higher connectivity, to better adapt to the development of urban UUS.

5. Conclusions

In this study, the AHP-CA coupling model was used to simulate and analyze the suitability evaluation of UUS development under the action of connection. Firstly, AHP is used to comprehensively consider the government, developers, and users, and a comprehensive evaluation index system of the suitability of UUS-connected development in the adjacent area of Erhai Basin is established. After calculating the weight ratio of each influencing factor, ArcGIS overlay analysis was used to evaluate the suitability of UUS. Then, according to the characteristics of the study area, the appropriate cell structure and CA operating rules were determined. Considering the inter-related development of adjacent UUS, the AHP evaluation results are adjusted to obtain the final evaluation results (Figure 11).

AHP results show that spatial location, POI mixing degree, and land use type are the most important influencing factors. Traffic conditions, road accessibility, and bus stop density are relatively less important. About 9.04% of the area (71.26 km²) is suitable for UUS development. These areas are mainly concentrated in the northwest corner of Erhai Valley, close to the regional political and economic center, most of the areas have considerable space resources and development value. These areas will be the focus of UUS development in the future. Through CA simulation and considering the development potential of the adjacent UUS connections, the final comprehensive evaluation result of the suitability of the adjacent UUS connections development in the Erhai Basin was obtained. The results show that the improvement area of suitability is about 10.73% (88.1 km²), mainly concentrated in the town center area. The development of UUS in these areas is critical to building a connected UUS-centric economy in the future.

This study simulated the suitability assessment of UUS development considering connected development to reflect the interaction between plots. In particular, it is of great significance for the city to formulate a forward-looking UUS development plan. In future studies, a complete comprehensive index system can be built by combining more abundant and comprehensive indicators, and different regions can be screened from the complete index system according to their characteristics. In addition, due to the small amount of UUS-related research data in the region, we cannot provide sufficient samples for the final evaluation of CA construction, so the accuracy of the research is limited.