Research on Evaluation of City–Industry Integration in Industrial Parks (2024)

1. Introduction

Today’s developed countries saw the rise of modern cities during industrialization, and this rise reflects wave-like progress in the coupling degree between industry and cities. In the post-industrialization period, some cities have begun the process of de-industrialization, and the degree of city–industry integration has declined significantly, with some cities trying to develop alternative industries to restore integration. For example, the decline of the textile industry in Napolis, which is in the “sunrise” region of economic growth in the United States, has deepened the degree of deindustrialization, harming employment rates and community development [1]; Cleveland in the “Rust Belt” of the United States, Bochum in the Ruhr area of Germany, and Halle in East Germany have experienced a gradual decline in heavy industry, resulting in the closure of numerous factories, abandonment of equipment, idle or inefficient use of land, and insufficient industrial support for the cities. Over the past 20 years, efforts have been made to develop alternative industries, such as education, biotechnology, medical research, healthcare centers, or culture, and some progress has been achieved [2,3,4] towards the recovery of city–industry integration. Developing countries are generally in the process of making the same wave-like progress in city–industry integration that developed countries have gone through, but whether they can complete this process remains an important challenge for their urban development.

As a direct carrier of industrialization and an engine of urban development, the issue of industry–city integration in industrial parks is one of the focal points of urban development. Industrial parks are important centers of urban employment and commodity production, and the establishment and construction of industrial parks is an effective strategy for promoting sustained and rapid economic and urban development and has been recognized and practiced worldwide, such as the Ulsan industrial park in Korea, Queensland industrial park in Australia, Burnside industrial park in Canada, Cikarang industrial parks in Indonesia, Tianjin Economic Development Area, and Taiwan Linhai industrial park in China, which play an irreplaceable and important role in regional economic and urban development [5,6,7]. However, some countries or regions’ industrial parks are facing serious city–industry mismatch. For example, some industrial parks in China have blindly paved “big stalls”, which occupy a large amount of territory, but have very low output [8], thus forming a kind of “empty city”. Meanwhile, some industrial parks only have manufacturing functions and lack service industry support, with an obvious dislocation between the residential and employment populations, thus becoming “isolated islands” where employees live a “pendulum-like” life. Some industrial parks also have more prominent environmental coupling coordination degree problems, which have become an obstacle for city–industry integration [9]. To this end, the National New Urbanization Plan requires “overall planning of production areas, living areas, commercial areas, and promoting functional mixing and city–industry integration” in the construction of new cities and new areas, and in recent years, China’s state leaders have proposed that it is necessary to achieve “city–industry integration, job-residence balance, ecological livability, and convenient transportation” in the construction of new suburban cities [10].

The evaluation of city–industry integration is an important part of the research on city–industry integration, and some scholars have evaluated the city–industry integration of some industrial parks in China from different latitudes and using different methods, and the following is a brief review of the main research results from the aspects of index system, index weight, evaluation scope, and evaluation results.

In terms of index system construction, Li, Y. and Zhang, Z. established an index system from three aspects: industrial development, urban construction, and people’s livelihoods [11]. Wang, X. [12] designed 18 second-level indicators and 77 third-level indicators to construct a framework system for the evaluation of city–industry integration in high-tech zones from three dimensions: an industrialization development index, an urbanization development index, and a separation coefficient for city–industry integration. Dong, W. [13] established three dimensions for the evaluation index system: spatial efficiency, operational efficiency, and industrial screening. Based on this, combined with the actual situation of each land parcel, six specific evaluation indicators were selected according to their importance and guidance, including floor area ratio, building density, plot size, usage, enterprise rating (A/B/C/D), and industry category. Tang, X.H. [14] constructed an evaluation index system for city–industry integration in development zones from the four dimensions of industrial development, population integration, spatial integration, and urban function. Jia, J. [15] selected 16 specific indicators from four aspects, industrial development, urban function, residents’ demands, and resource allocation, to build an index system. The Evaluation Rules for Intensive Use of Land in Development Zones (trial version in 2014) issued by the Ministry of Land and Resources puts forward the evaluation criteria for intensive use of land in city–industry integrated industrial parks. In addition, indicators such as resource allocation, social development [16], spatial integration [17], natural environment [18], basic social services [19], and scientific and technological innovation [20] have been emphasized by different scholars.

There are also some other academic achievements in industrial park research that provide logical support for the inclusion of air pollution, land use efficiency, and other indicators in the evaluation of city–industry integration in industrial park, which are briefly summarized as follows. The development of industrial parks is conducive to attracting investment, raising tax revenues, and increasing employment, thus enhancing the economic efficiency of industrial parks and the regions in which they are located [21]. However, mixed industrial parks have a major environmental and spatial impact [22], and they tend to burn large amounts of fuel and produce greenhouse gases [23], which can negatively affect the prices of neighboring residences [24]. Therefore, environmental performance was key to the success of eco-friendly industrial parks [25] and it is necessary to implement sustainable eco-industrial development (EID) strategies [26] and researchers should choose indicators from economic, social, and environmental aspects to evaluate the development performance of eco-industrial parks [27]. As a matter of fact, both developed and developing countries have tried to make efforts to promote the ecologization of industrial parks and have successfully built some eco-industrial parks [28]. For example, after 50 years of industrial development, South Korea started to promote the ecologization of industrial parks in 2005 and has achieved some notable successes [29].

As for the method of evaluation, the entropy value method [30,31,32,33], analytic hierarchy process [14,34,35], coupled coordination relationship model [36,37], factor clustering analysis [11,38,39], gray association model [17], Delphi method [13], double aggregation mechanism [40], fuzzy hierarchy [37], four-grid quadrant method [41], as well as another method combined with the expert scoring method [20] were mainly adopted to implement the evaluation.

Regarding the scope of the evaluation, Tang, X.H. [17] covered five representative development zones in Shanghai; Jia, J. [15] covered seven national high-tech industrial development zones in Henan Province; Gan, L. [42] covered 18 prefecture-level cities in Sichuan Province; and the research of Li, Y. [43] covered 31 regions (including provinces, autonomous regions, and municipalities directly under the Central Government) in China. Wang, X. [12] covered 56 national high-tech zones; Zheng, J.Q. [44] picked Xiamen Tongxiang high New Town as the object of evaluation; Li, Y. [45] covered nine prefecture-level cities in Fujian Province; Kong, X. [46] covered 13 administrative districts in Wuhan Province; and Luo, D. [47] covered 21 cities in Guangdong Province.

As for the evaluation results of the overall development of city–industry integration, Zou, D.L. [48] and Li, Y. [43] believed that the overall level of city–industry integration development in China is not high at present, and the overall characteristics of a high level of development in the east and a low level of development in the west are very significant. In Sichuan, a large western province of China, the level of city–industry integration is not high, and the degree of city–industry integration between regions is also unbalanced [30].

Scholars have made many advancements in the development of index construction, index weight establishment, evaluation implementation, and countermeasures and suggestions, but the evaluation of city–industry integration in development zones is relatively weak on the whole, and some existing research results are also worth further consideration.

In terms of the scope of the evaluation, the concept of city–industry integration was first proposed in Xinxiang City, Henan Province. Later, local governments at all levels, as the promoters of city–industry integration, approached city–industry integration from the perspective of urban areas and applied it to new development areas, such as economic development zones. However, many scholars view city–industry integration from the perspective of prefecture-level cities in municipal areas or even provincial regions, interpreting it as the coordination of industrialization and urbanization, which extends beyond the scope of new urban areas. At present, most evaluations of city–industry integration are based on the above understanding. However, due to the limited availability of urban data and for other reasons, there are relatively few evaluations of city–industry integration in new urban areas such as development zones that are based on the original understanding of the concept of city–industry integration, so it is necessary to strengthen it.

In the construction of the indicator system, existing research has been carried out based on a narrow understanding of industry–city integration. When constructing the evaluation index system for city–industry integration, existing research is based on the following understanding of city–industry integration: it is the spatial integration of industry, population, and urban carriers, that is, the spatial proximity of urban production function areas and living function areas and their mutual support. This state is indeed the integration of industry and city, however, the core essence of the required “spatial proximity” is to make workers’ commutes more convenient and ensure the smooth flow of transportation. For industrial parks with industrial production pollution, if the layouts of industrial production functional areas and living functional areas are still in accordance with the principle of spatial proximity, this will lead to city–industry conflicts; if the two are moderately separated, but can satisfy the above core elements, they should be regarded as having achieved successful city–industry integration. Based on this understanding, constructing an indicator system for the industry–city integration of industrial parks and implementing evaluations are research gaps that currently exist.

Furthermore, some scholars select too many indicators, in which there is inevitably duplication and intersection, and it is easy to inappropriately increase the weight of individual indicators. This in turn increases the difficulty of obtaining data and reduces the feasibility of an evaluation; some scholars, in addition to the inclusion of a number of indicators that can directly reflect the effectiveness of city–industry integration, have also selected a larger number of factors that have not been proven to be causal by conclusive evidence, and some of the chosen indicators do not have a close and direct causal relationship with urban integration.

Evaluating city–industry integration in development zones is an important part of the process to understand their current situations, problems, and experiences. We intend to clarify the original meaning of city–industry integration, construct an evaluation index system, carry out an evaluation of city–industry integration based on the sample data of national economic and technological development zones, and summarize the model of city–industry integration in industrial development zones so it can be used as a reference in future works.

2. Evaluation Indicator System of City–Industry Integration in Industrial Parks

From the perspective of urban areas rather than prefecture-level cities or provincial administrative districts, city–industry integration should be understood as a coordinated, balanced, rationally laid out and mutually supportive relationship within and among production and service functional areas centered on catering to the urban population (this includes members of the agricultural migrant population who work and live in the city and the members of the population who work and live in the suburbs of the city). Based on the basic characteristics of city–industry integration in industrial parks, we plan to construct an evaluation index system mainly using the two following aspects: land–industry integration and residence–industry integration (refer to Table 1).

2.1. The Original Meaning of Industry-City Integration

Understanding city–industry integration from the perspective of urban areas and applying it to new urban areas is the primary task of researching and promoting city–industry integration. A search on the website of the Central Government of China and CNKI found that the concept of city–industry integration was first proposed in Xinxiang City, Henan Province, in February 2010 to express the city’s requirements for the construction of the “Pingyuan New Area” (please refer to the Appendix A). Subsequently, local governments at all levels began to use this concept, applying it primarily to industrial parks, development zones, and other new urban areas. The requirement is that these new areas are not turned into “islands” or “empty cities”, but instead are built into comprehensive and functional urban areas capable of maintaining production, life, and ecological functions. For its part, China’s central government first used the concept in the National New Urbanization Plan issued in 2014 to clarify its requirements for the construction of new urban areas. City–industry integration is an applied concept promoted by the government, and the academic research on city–industry integration should facilitate the government’s promotion of city–industry integration in new urban areas. Some scholars understand city–industry integration from a more macro perspective and regard it as the benign interaction between industrialization and urbanization in cities, provinces, and even the whole country. Although the current understanding of city–industry integration is not wrong, the essential first step is to understand, study, and promote industrial parks and other new urban districts based on the initial meaning of the concept.

From the perspective of urban areas, city–industry integration should be understood as a coordinated, rationally laid out, and mutually supportive relationship within and between production and service functional areas centered on the urban population (see Figure 1). There are three basic elements in a city, namely industry, population, and carrier. Urban industry mainly includes secondary and tertiary industries, population mainly includes the population engaged in the industry (working or retired) and their dependents, and urban carriers include different uses of urban land and the buildings, structures, plants, and other attachments found within the area. The three basic elements of the city interact to form the three major functional areas of the city centered around the population: the production functional area, the service functional area, and the ecological functional area. The production functional area contains the secondary industry and its urban carriers (the land and its attachments) and the service functional area contains the tertiary industry and its carriers.

Production functional areas are usually various forms of industrial parks, within which the supply of industrial land, industrial buildings, or structures needs to be coordinated with industrial development (the urban production industries also include construction, but since land for construction is not permanently available, this paper only refers to industry when discussing urban production function areas), if there is too much industrial land, there will be idle or inefficient use of industrial land, and if the supply is insufficient, it will restrict industrial development.

The service functional area provides productive services, life services, and public services and the service sector and its carriers are required to be coordinated and mutually supportive. For example, the life service sector’s carriers mainly include housing, commerce, entertainment, and recreation, as well as other aspects of the land and its buildings and structures. If housing exceeds the living requirements of the urban population, there may be vacant houses. In some places, there is a serious oversupply of housing in new urban areas and insufficient population and industrial agglomeration, and these areas have been referred to by the media as “empty cities” or even “ghost cities”. The productive service sector’s carriers mainly include commercial land and associated buildings (office buildings, etc.) as well as logistics, warehousing land, and associated buildings (warehouses, storage yards, etc.). If the supply of the business carriers exceeds the needs of the development of the business service sector, the office buildings are inefficiently utilized or even vacant in large areas (these areas are also referred to by the media as “empty cities”), and the business income per unit area is reduced.

In terms of the relationship between the production functional area and the service functional area, on the one hand, the production functional area supplies material products to the service functional area, and promotes the development of the service functional area, forming the basic driving force of urban development, at the same time generating a certain amount of pollution [49,50], which may have a certain negative impact on the service functional area. On the other hand, the service functional area provides support for the development of the production functional area, including providing residential, commercial, medical, educational, pension, entertainment, and other life services for production and employment personnel and their dependents, and providing enterprises with R&D, consulting, logistics, human resources, finance, and other productive services.

Rational spatial layout—the desirable distance between the production and service functional areas, which is determined by balancing the benefits of reducing pollution with the costs of transportation—is very important. Most areas of China have not yet completed industrialization and production functional areas (essentially various forms of industrial parks) have traditional manufacturing as their leading industry, causing there to be certain industrial pollutants, such as air, noise, dust, sewage, and there are even certain potential safety hazards; therefore, the service functional areas tend to maintain a certain spatial distance from such general production functional areas, and the farther the distance, the more they are able to raise their pollution avoidance utility, and if the distance is far enough to completely avoid the pollution of the production functional area, the pollution avoidance utility will no longer increase. Evasion utility will decrease as the distance increases, meaning its marginal utility is decreasing. For example, when the distance between the service functional area and the general production functional area is 0, the marginal utility of pollution avoidance 1 km from the layout of the service functional area is very high, and when there is a greater distance between the service functional area and the production functional area, the utility of pollution avoidance achieved by being further than 1 km away from the area will be lower than the former. So, we can see a pollution avoidance marginal utility curve decreasing with distance in Figure 2. On the other hand, as the distance between the service function area and the production function area increases, the transportation cost of the latter to obtain the services of the former increases. In reality, this kind of transportation cost is manifested as the commuting cost paid by the employees of the production functional area to obtain the residential services of the service functional area, the transportation cost paid by the residents of the production functional area to obtain the commercial, entertainment, educational, medical, and other life services or public services of the service functional area, and the transportation cost that the enterprises in the production functional area need to pay to obtain the productive services and public services of the service functional area. In general, people’s pain level in traffic will accelerate with the increase in distance, and self-driving traffic will even accelerate fatigue levels with the increase in distance and then accelerate accident probability, so Figure 2 presents a marginal cost curve for transportation that increases with distance. As shown in Figure 2, according to the optimization principle that marginal utility and marginal cost are equal, we can see that the desirable distance between the production and service functional areas is D₁ points.

The green manufacturing system helps to promote the integration of industry and cities. If a city has finished the industrialization process, the production function area is transformed and upgraded into a high-tech green industrial park, which generates very little pollution externally; thus, the marginal utility value of the pollution specification on the boundary of the production functional area is very low, and the decreasing range with the increase in distance is very small. In this case, the production function area and the service functional area should be adjacent to each other or even in a mixed layout; for example, as with the situation presented by point D₂ in Figure 2. At this time, if the service functional areas such as residential areas are far away from the industrial park, the employees of the industrial park may bear higher transportation costs compared to value of the pollution avoidance utility, and the degree of urban traffic congestion will be exacerbated, which will deviate from the requirements of city–industry integration.

In cases where a green manufacturing system has not yet been established, some scholars and officials will have a one-sided understanding of city–industry integration and they may think that integrating production functional areas and service functional areas—such as in point D₂ in Figure 2—constitutes city–industry integration. In fact, under the condition that there is obvious pollution in the production function area and the marginal utility curve of pollution avoidance remains unchanged, the welfare level represented by the D₂ point is far lower than the D₁ point, and the people living there face prominent health and safety threats, so that the production functional area and the service functional area cannot support each other, which is contrary to city–industry integration.

Rail transit is an important component in the coordination and integration of production and service functional areas. If the subway, light rail, and other rail transit systems are built between the production functional area and the service functional area, the increased comfort can lead to a marginal transport cost curve that is lower than the general marginal transport cost and grows more slowly, and the desirable distance between the service functional area and the production functional area can be increased to D3 (shown in Figure 2) under the condition that the pollution level of the production function area remains unchanged. In this position, the service function area will be less affected by the pollution of the production function area at a lower transportation cost, and the employed population and enterprises in the production function area can enjoy the living and production services of the mother city by paying a lower transportation cost, or the transportation cost between the two can be greatly reduced without changing the original distance D₁ between the production function area and the service function area.

In short, city–industry integration should be understood as a coordinated, rationally laid out, and mutually supportive relationship within and between production and service function areas centered around the urban population, with features that include an intensive and efficient production function area and an intensive and efficient service function area, which are measured using the indicator of “land–industry integration” in the following sections. The features also include the spatial layout of production function areas and service function areas, which should comply with the requirements of pollution avoidance and reduce the transportation cost between them through rail transit and other means, which is measured by the indicator of “residence–industry integration” in the following sections.

2.2. Indicators of Land–Industry Integration in Industrial Park

Land–industry integration in an industrial park implies highly intensive land use in the production functional zone, wherein industry is integrated with its carrier—industrial land—and the service sector is integrated with its carrier—commercial service land—thereby achieving optimal efficiency in both industrial and service land utilization. The first-level indicator of land–industry integration includes two second-level indicators, industrial land efficiency and service land efficiency, among which the industrial land efficiency indicator includes three third-level indicators, industrial investment intensity, employment density, and output intensity. The industrial output per mu (mu is a unit of area (=0.0667 hectares)) is the most direct index to measure the efficiency of industrial land, and the investment intensity of industrial land is an important index affecting the output intensity of industrial land. In view of the fact that China is a populous country, the employment issue has prominent social significance. Although industry is moving in the direction of intelligence services, the employment density on industrial land is still regarded as one of the indicators to measure the efficiency of industrial land.

To calculate the investment intensity of industrial land, since complete data on fixed asset investments are lacking, the weighted average of registered capital and paid-in capital of unit industrial land is used, in which the weight of the registered capital and paid-in capital each accounts for 50%, and the area of the industrial land is measured by the area of the factory area picked up on the map (the same applies below). While this measurement method may not be entirely accurate, it effectively reflects the actual land utilization status for industrial development in each industrial park. In terms of the calculation method for the employment density index, due to the lack of complete employment data, it is measured by the number of industrial enterprise employees paying social security on each unit of industrial land in the industrial park, which will lead to an underestimation of the actual employment density. However, it is acceptable to assume that the degree of underestimation of the industrial employment density in each industrial park is roughly the same. Therefore, the industrial employment density of each industrial park measured in this way is comparable.

The development of many industrial parks not only encompasses the growth of several industries, but also fosters the advancement of the service sector. Another second-level indicator for integrating industry with urban land is how efficiently land is utilized for services within industrial parks. This indicator is comprised of two three-level indicators: service output intensity and service employment density. The output intensity of the land used for the service sector is measured by the added value per unit of urban land occupied by the service sector, in which “land for service sector in industrial park = built-up area of industrial park—factory area—green space and water area—residential area”. The land area for the service sector in industrial parks determined in this way is also rough. However, we are able to compare industrial parks using the same calculation method. The employment density of the service sector is measured by the number of service enterprise employees in the industrial park participating in social security ÷ the land area for the service sector in the industrial park. We do not take the service sector investment intensity as the evaluation index, because most service sector enterprises adopt an asset-light mode of operation, and the impact of the service sector’s investment intensity on enterprise operation performance and land use efficiency is not prominent.

2.3. Indicators of Residence–Industry Integration in Industrial Park

The residence–industry integration in industrial parks refers to the coordination and integration of industrial production functional zones with residential areas. This includes two secondary indicators: the matching degree between residence and environment and supportive rail transit facilities. The matching degree between residence and environment refers to the degree of harmony between the residential area in the industrial park and its external air environment, whose value is equal to “1 − |z_r + z_a − 1|”, where z_r refers to the supportive residential score of the industrial park, which is equal to “the standardized value of per capita residential area (range normalization, the value is 0–1, the same as below) × 0.5 + the standardized value of the proportion of residential area to built-up area × 0.5” and the value is in the range of [0, 1].

z_a stands for the comprehensive air quality index of the industrial park, whose formula is as follows:

“AQI standardized value of an industrial park × 0.5 + standardized value of (AQI of an industrial park ÷ AQI of the mother city of the industrial park) × 0.5”.

The air quality composite index z_a of the industrial park serves as a reverse indicator, reflecting a portion of the impact exerted by ETDZs on their surrounding environment. A smaller value indicates that the air quality is better, both in absolute terms and relative to the average air quality of the city where it is situated, rendering it more conducive for habitation; on the contrary, a higher value suggests that livability is weak.

The match score between residence and environment, whose value is in the range of [0, 1], denoted as “1 − |z_r + z_a − 1|”, indicates a higher level of coordination when the value approaches 1. Conversely, when it approaches 0, both z_r and z_a are either small or large simultaneously, indicating a low level of environmental coordination between residential and industrial areas. (1) When both z_r and z_a achieve a high value of 1 at the same time, “1 − |z_r + z_a − 1|” reaches a minimum value of 0. This implies that the air quality score of this industrial park based on the level value and the value relative to the mother city is extremely poor in the sample industrial parks. However, the residential supporting score based on the proportion of supporting residential areas in the built-up area of the industrial park and the per capita living space is extremely high, thus a large number of residential spaces are affected by the negative externalities of industrial development, and the residential areas is not in harmony with the environment. (2) When both z_r and z_a achieve a low value close to 0 at the same time, “1 − |z_r + z_a − 1|” still reaches a minimum value of 0. This indicates that the air quality score based on the level value and the value relative home city is very high in the sample industrial parks. However, there is only a minimal percentage of space allocated for supportive residential areas within these industrial parks, or per capita living space within these parks is very limited. Consequently, the industrial park has not provided enough space for residents’ needs. That is to say, the environment is suitable for configuration but there is not enough residential space to meet the living needs of member of the industrial population and their dependents, which leads to an impediment to industrial development and an imbalance between production and housing.

If the match score between residence and environment (1 − |z_r + z_a − 1|) is close to 1, then z_r and z_a change in the opposite direction, representing a high match degree between residence and environment: (1) as z_a approaches 0 and z_r approaches 1, it implies that the air quality in the industrial park is much better compared to the overall air quality of the city. The residential area within the industrial park constitutes a substantial proportion of its built area, providing abundant per capita living space. This positive interaction between industrial development and supportive residential facilities ensures a harmonious living environment within the industrial park. (2) As z_a approaches 1 and z_r approaches 0, it indicates that the air quality score of this industrial park is extremely poor, based on the value relative to the mother city. Accordingly, the residential area within the industrial park accounts for an insignificant portion of its built area and the per capita living space is very small. Consequently, residents seek residence far away from the industrial park to minimize negative externalities caused by industrial development on their living spaces, which is an important form of matching between residence and environment. (3) As long as the sum of the supporting score of the residential area and the score of the comprehensive air quality index is close to 1 (i.e., z_r + z_a→1), they all belong to the coordinated state between the residence and environment.

3. Data Collection and Processing of City–Industry Integration in Industrial Parks

According to the principles of necessity and feasibility, we have selected several national industrial parks—ETDZs—to conduct an evaluation on city–industry integration. This evaluation aims to reveal advanced experiences as well as possible shortcomings in the city–industry integration of these industrial parks. Simultaneously, the evaluation process and results will also test the rationality of the evaluation index system and whether it can be widely used.

3.1. Data Collection Methods and Procedures

Collecting data for the evaluation is a fundamental part of the evaluation process and includes seven steps, as shown in Figure 3.

3.1.1. Select the Evaluation Objects of City–Industry Integration in ETDZs

We first selected 23 ETDZs as the evaluation objects for city–industry integration in industrial parks according to the principles of typicality and feasibility. The specific selection criteria are as follows:

(1): Inclusion in China Development Zone Yearbook 2020 Edition;
(2): Economic growth data are disclosed;
(3): The planning map can be found through open channels.

Eventually, we selected 23 national ETDZs (see Figure 4), including Wuhan, Beijing, Chengdu (in view of data availability and matching, only the Damian Street area of Chengdu ETDZ is included in the evaluation. This area is the core area of Chengdu ETDZ and Chengdu ETDZ only refers to the Damian Street area in this paper), Ningbo Daxie, Zouping, Longyan, Jiashan, Hanzhong, and Quanzhou ETDZs, and collected their GDP data.

3.1.2. Determine the Scope of the Sample ETDZs on the Map

The planning map of ETDZs presents the geographical scope of the ETDZs. We obtained the planning maps of each ETDZ through open channels and compared them to the development zone boundaries published on the website of the Ministry of Natural Resources to confirm the validity of the planning maps. Amap updates its geographic information every six months (the source of the information is Amap customer service); this ensures the information is highly current. Therefore, we primarily utilized Amap as our source for the geographic information required for this research project. Based on the determined geographical extents from the planning maps, we obtained the vertex coordinates of the boundary of the sample ETDZs on Amap.

3.1.3. Pick up the Polygon Vertex Coordinates of Factories, Green Space Water Area and Unbuilt Area in ETDZs and Calculate Their Area

Firstly, we extracted the polygon vertex coordinates of factories, green areas, and water areas within ETDZs using Amap. Subsequently, we employed an Excel editing formula to calculate the respective polygon areas in batches. While collecting factory coordinates, we discerned factory areas based on building features depicted in satellite maps while excluding non-factory regions such as logistics areas and shopping centers. Moreover, when acquiring coordinates for unbuilt-up areas, contiguous vacant lands that had undergone urban infrastructure development like roads and underground pipe networks were also considered to be unbuilt-up territories.

3.1.4. Obtain the Data of Enterprises in ETDZs

Based on the boundary information of the ETDZs, we used the Qianxun map program on the Qichacha website to pinpoint the exact location of the ETDZs. This allowed us to identify all enterprises within the ETDZs and obtain micro data, such as registered capital, paid-in capital (paid-in capital data from the Tianyancha website), and the number of insured people, and then we used the Baiteng network to check all of the enterprises’ authorized invention patent data by searching each enterprise’s name.

3.1.5. Get Data on the Land Area of Residential Areas in ETDZs

The per capita residential floor area is suitable for measuring the supportive residential area, but there are many real estate projects in ETDZs that cannot be searched on the Internet, so we use the land area of residential areas to measure the supportive residential areas in the ETDZs. Residential areas refer to areas enclosed by outer roads. This is similar to the actual area of land used for residential purposes. At the same time, we used the XiaoO map 0.9.3.0 software to obtain the name, type, and other basic information of all the houses in the development zone, and then obtained the residential area, building area, and floor area ratio of the residential district from real estate industry information websites such as Fangtianxia, Lianjia, and Anjuke. As for the old residential areas where the land area data are missing, we obtained the data through calculations after picking up the coordinates.

3.1.6. Get Air Quality Index (AQI) Data

We obtained the air quality index (AQI) for 2021 from air monitoring stations within the ETDZ and from all monitoring stations in the cities where the ETDZs are located using the Ture Air website (https://www.zq12369.com/, accessed on 23 May 2024). For monitoring stations in ETDZs with only one air quality monitoring point, the AQI of the ETDZ was used to measure the air quality of the ETDZ. For monitoring stations in ETDZs with more than one air quality monitoring point, the average AQI was adopted. For monitoring stations in ETDZs with no air quality monitoring points, the data from the nearest air quality monitoring point were adopted. For the air quality of the mother city of the ETDZs, the average AQI of all the monitoring points in the city were used.

3.1.7. Obtain the Data of Rail Transit Stations

We used the interest point query function of XiaoO map to draw polygons according to the boundary vertices of each ETDZ, queried the rail transit stations inside the polygon, and exported the data to analyze the rail transit stations in the ETDZs.

3.2. Calculation Process and Data Characteristics of the Complex Index—Matching Degree between Residence and Environment

The sub-index for the intensity of the production functional area is calculated using the formula in Table 1, which is relatively simple. The calculation of the matching degree between residence and environment for the residence–industry integration index is more complicated.

3.2.1. Supportive Residential Area Score in Industrial Parks

As mentioned above, the matching degree between residence and environment is determined by the supportive residential area (z_r) and the air quality composite index (z_a) (refer to Table 1); the supportive residential area score z_r is calculated according to the formula “supportive residential area score = standardized value of the proportion of residential areas in built-up areas × 0.5 + standardized value of per capita residential area × 0.5”. For the sample industrial parks, the z_r value for the Kunming ETDZ is a perfect score (the possible reason for this lies in the district’s integrated management, where the residential area of these ETDZs is planned based on the total permanent population of the administrative region, along with other factors, and only workers employed by enterprises in the ETDZ are used as the denominator, which will lead to an abnormally high per capita residential area). Meanwhile, although the Beijing ETDZ’s ranks fifth among the sample ETDZs in terms of its proportion of residential area to built-up area, its high employment density leads to a relatively low per capita residential area, placing its z_r at the lower–middle level.

3.2.2. The Inverse Matching Relationship between the Air Quality Composite Index and Supportive Residential Area

As mentioned above, the air quality composite index z_a of the ETDZs is utilized to reflect the impacts to the surrounding environment caused by these zones. It is calculated as follows: “standardized value of ETDZ AQI × 0.5 + standardized value of (AQI of ETDZ ÷ AQI of the mother city of ETDZ) × 0.5”. The standardized value of AQI in ETDZs reflects the absolute number of the air quality in ETDZs, while “AQI of ETDZ ÷ AQI of mother city of ETDZ” reflects the relative number of the air quality in ETDZs. If both render a high value simultaneously, then the z_a value is higher, indicating heavy air pollution within ETDZs, with worse air quality compared to the entire city. This provides sufficient evidence of urban air pollution caused by ETDZs. Therefore, residential areas should not be planned in ETDZs, but should be arranged in the main urban area far away from the ETDZs, meaning that the z_r value should be low. If both values are low, then the z_a value is small, indicating that the air quality is high (≤50) or close to high quality in ETDZs and better compared to the overall air quality level of the entire city. For example, the Beijing and Daya Bay ETDZs, demonstrate that ETDZs can have minimal impact on urban air quality and it is advisable to plan for a sufficient scale of residential areas whenever possible within ETDZs, meaning that the z_r value should be higher. If the absolute AQI in an ETDZ is high, but the relative AQI is low, then the z_a value is in the middle, indicating that the ETDZ has weak livability but is overall better than the level of livability in the city. Therefore, a medium-scale residential area should be laid out within the ETDZ, and a certain amount of the residential area can be laid out within the main urban area or other suitable areas, meaning that the z_r value should be in the middle. If the absolute AQI in an ETDZ is low, but the relative AQI is high, and then the z_a value is in the middle as well, this suggests that the air quality in the ETDZ is good, but worse than in the whole city to some extent. In such cases, it is suitable to lay out a medium-scale residential area with better livability both within the ETDZ and within the main urban area so as to ensure people’s pursuit of a better life, meaning that the z_r value should be in the middle.

The negative correlation between the air quality composite index z_a and supportive residential area score z_r is shown in Figure 5, and the relationship between z_a and z_r is fully reflected in Korla, Zouping, Rugao, Hai’an, Lianyungang, Chengdu, Changchun, and other ETDZs. This negative correlation is not obvious in some ETDZs, indicating that the concept of city–industry integration has not been fully implemented in terms of residential support due to subjective or objective factors.

3.2.3. Matching Degree Score between Residence and Environment in Industrial Parks

Based on the z_r and z_a values, the score of the matching degree between residence and environment of the ETDZ is calculated, as shown in Figure 6. The match score between residence and environment is in the range of [0–1], and z_r is negatively correlated with z_a, and the closer the matching degree between residence and environment (1 − |z_r + z_a − 1|) is to 1, the more obvious the negative correlation between the air quality composite index and the per capita supporting residential area scale in the ETDZ, and the more coordinated the air quality and residential facilities in the ETDZ, the more conducive it is to promoting the residence–industry integration. Figure 6 shows that the matching scores between residence and environment in industrial parks such as the Hai’an, Lianyungang, and Chengdu ETDZs are at the forefront. Among them, the Haian, Rugao, and Korla ETDZs scored higher in the matching degree between residence and environment because of the poor air quality in the ETDZs (high z_a) and the low supporting residential scale (z_r) in the ETDZs. The reason for the high score of the matching degree between residence and environment in the Lianyungang and Jiashan ETDZ is that the comprehensive air quality of the ETDZ is intermediate to low level (medium-high z_a), and the supporting residential area scale (z_r) in the ETDZ is at a low to medium level too. While the reason why the matching degree between residence and environment in Chengdu ETDZ is in the forefront is that the comprehensive air quality of the ETDZ is at a medium to high level (medium-low z_a), and the supporting residential scale (z_r) in the ETDZ is at a medium to high level too.

On the other hand, the matching degree between residence and environment in Ningbo Daxie, Beijing, Longyan, Ningguo, and Linyi ETDZ lags behind. Among them, the backwardness of the Ningbo Daxie, Beijing, Longyan, and Ningguo ETDZs can be attributed to their high comprehensive air quality (low z_a) and the insufficient scale of the supporting residential area within the ETDZs (low z_r), which is suitable for the layout of more residential space. The backwardness of the match degree between residence and environment in the Linyi ETDZ is due to the low comprehensive air quality (high z_a) and the high scale of the supporting residential area in the ETDZ (high z_r), which makes the larger-scale residential areas suffer from the impact of air pollution in the ETDZ.

4. Evaluation Process and Results of City–Industry Integration in Industrial Parks

4.1. Standardization of Basic Indicators

Because the original data dimensions and orders of magnitude of the selected indicators are different, the original data should be standardized first for comprehensive evaluation and comparison, and we use the range standardization method to complete this work. For the original indicator with a positive nature, the calculation formula is $\frac{X - X_{m i n}}{X_{m a x} - X_{m i n}}$ . For the negative indicators, such as the supportive residential area score, the calculation formula is $\frac{X_{m a x} - X}{X_{m a x} - X_{m i n}}$ . After range standardization, all the indicator values are between 0 and 1, and all are positive indicators. The range standardization values of each indicator are shown in Table 2.

4.2. Determine the Weight of Indicators Based on the Analytic Hierarchy Process of Expert Scores

We use the Analytic Hierarchy Process (AHP) combined with expert scoring to determine the weights of each evaluation indicator, which is a hierarchical weight decision analysis method proposed by American operations researcher Professor T.L. Satty. The analytic hierarchy process of expert scoring can avoid the Matthew effect in the process of establishing the weight of the evaluation indicator and can give higher weights to the evaluation indicator for city–industry integration in industry parks, which better reflects the new development concept, can better meet the requirements of high-quality development, and can better meet the needs of the people. Of course, there is a certain subjectivity in this method, and to overcome its effect, the number of experts we solicited reaches 18. In addition, we will also use the entropy weighting method, which is fully objectively empowered, in the fourth section of this chapter to test the robustness of the AHP evaluation results.

4.2.1. Modelling the Hierarchy

Our work using the analytic hierarchy process includes five steps: establishing a hierarchical model, constructing a comparative judgment matrix, ranking a single hierarchy, checking consistency, and determining the weight of the indicators (since there is no need to uniformly compare the weights of the fourth-level indicators in this indicator system, and the same is true for the third-level indicators, there is no need to perform the common hierarchical total ranking and consistency test in the analytic hierarchy process), and the indicator system constructed above (refer to Table 1) can correspond to the first step, the hierarchical model. The overall goal A of the highest level is city–industry integration. The second level is the criterion layer for evaluating the overall objectives, including two criteria: land–industry integration (B1) and residence–industry integration (B2). The third layer is the sub-criterion layer of the evaluation criterion layer, which is used to evaluate the industrial land efficiency (C11) and service land efficiency (C12) of B1, and the matching degree of residence and environment (C21) and rail transit support (C22) of B2, for a total of four sub-criteria. The fourth layer is the element layer that measures the sub-criterion layer, including four elements consisting of the investment intensity (D111), employment density (D112), density of invention patents on industrial land (D113), and industrial output intensity (D114) for measuring C11, and three elements consisting of the service sector output intensity (D121), density of invention patents on service land (D122), and service sector employment density (D123) for measuring C12.

4.2.2. Constructing the Comparison Discriminant Matrix

Based on the analytic hierarchy process, we designed a questionnaire based on the weighting of the evaluation indicator for industry–city integration in industrial parks (see Appendix B) and invited 18 urban economic experts from academia and various industries to fill in the questionnaire, asking them to compare and score each variable according to their own knowledge and experience, from the criterion layer (B layer) down to the factor layer (D layer), according to the importance of each variable in the evaluation of the variables in the previous layer, and the measurement standard was implemented according to the two-pair comparison scale table. The questionnaire was fully recovered, and each expert’s score formed five judgment matrices: A–B, B1–C1, B2–C2, C11–D11, and C12–D12. Among these matrices, the A–B matrix presents the weights of the two variables of the criterion layer in comparison with each other, land–industry integration (B1), and residence–industry integration (B2); the B1–C1 and B2–C2 matrices present the weights of the variables of the sub-criterion layer (C) compared with each other in the evaluation of the variables of the upper layer (criterion layer B); and the C11–D11, C12–D12 matrices present the weights of the variables of the element layer (D) compared with each other in the evaluation of the variables of the upper layer (sub-criterion layer C). Due to the limited space, the discriminant matrix formed by the scores of the 18 experts is not fully presented. Taking the first expert as an example, the five discriminant matrices formed according to the questionnaire scores are as follows:

$A – B_{1} : \begin{matrix} B 1 & B 2 \\ B 1 & 1 & 1 / 2 \\ B 2 & 2 & 1 \end{matrix} B 1 – C 1_{1} : \begin{matrix} C 11 & C 12 \\ C 11 & 1 & 2 \\ C 12 & 1 / 2 & 1 \end{matrix} B 2 – C 2_{1} : \begin{matrix} C 21 & C 22 \\ C 21 & 1 & 2 \\ C 22 & 1 / 2 & 1 \end{matrix} C 11 – D 11_{1} : \begin{matrix} D 111 & D 112 & D 113 & D 114 \\ D 111 & 1 & 3 & 4 & 1 / 5 \\ D 112 & 1 / 3 & 1 & 2 & 1 / 7 \\ D 113 & 1 / 4 & 1 / 2 & 1 & 1 / 8 \\ D 114 & 5 & 7 & 8 & 1 \end{matrix} C 12 - D 12_{1} : \begin{matrix} D 121 & D 122 & D 123 \\ D 121 & 1 & 3 & 2 \\ D 122 & 1 / 3 & 1 & 12 \\ D 123 & 1 / 2 & 2 & 1 \end{matrix}$

Regarding the specific meaning of the weights in the determination matrix, taking the C11–D11₁ matrix as an example, when measuring the industrial land efficiency (C11), the industrial investment intensity (D111) is slightly more important than the industrial employment density (D112), and investment intensity (D111) is between slightly and significantly more important than the density of invention patents on industrial land (D113), and the industrial output intensity (D114) is significantly more important than the industrial investment intensity (D111) and is more important than the industrial employment density (D112).

4.2.3. Hierarchical Single Sorting with Individual Expert Weights for Indicators

Hierarchical single sorting is the process of finding the individual weights of the experts of each variable in the previous layer of variables according to the weights in the judgment matrix. We use the “sum method” for our calculations, and the process includes three steps:

In the first step, the column vector of the judgment matrix is normalized, that is:

${\tilde{M}}_{k} = (\frac{a_{i j k}}{\sum_{i = 1}^{n} a_{i j k}})$

(1)

where a_ijk is the weight reflecting the importance of factor i relative to factor j based on the score of the k-th expert; in this case, the maximum value of k is 18, the maximum value of i and j in the C11–D11 and C12–D12 matrices is 4 and 3, respectively, and the maximum value of i and j in the rest of the matrices is 2.

In the second step, $\tilde{M}$ is summed by rows:

${\tilde{W}}_{k} = {(\sum_{j = 1}^{n} \frac{a_{1 j k}}{\sum_{i = 1}^{n} a_{i j k}}, \dots, \sum_{j = 1}^{n} \frac{a_{n j k}}{\sum_{i = 1}^{n} a_{i j k}})}^{T}$

(2)

In the third step, $\tilde{W}$ is normalised in a similar way to the first step to obtain the weight vector W_k = (w_1k, ..., w_nk)^T

$C 11 - D 1 1_{1} = [\begin{matrix} 1 & 3 & 4 & 1 / 5 \\ 1 / 3 & 1 & 2 & 1 / 7 \\ 1 / 4 & 1 / 2 & 1 & 1 / 8 \\ 5 & 7 & 8 & 1 \end{matrix}] \overset{(1)}{\to} {\tilde{M}}_{1} = [\begin{matrix} 0.152 & 0.261 & 0.267 & 0.136 \\ 0.051 & 0.087 & 0.133 & 0.097 \\ 0.04 & 0.044 & 0.067 & 0.085 \\ 0.760 & 0.609 & 0.533 & 0.681 \end{matrix}] \overset{(2)}{\to} {\tilde{W}}_{1} = [\begin{matrix} 0.816 \\ 0.368 \\ 0.233 \\ 2.583 \end{matrix}] \overset{normalization}{\to} W_{1} = [\begin{matrix} 0.204 \\ 0.092 \\ 0.058 \\ 0.646 \end{matrix}]$

(3)

The above W₁ matrix shows that when evaluating industrial land efficiency (C11), the individual expert weights for investment intensity (D111) are employment intensity (D112), density of invention patents on industrial land (D113), and industrial output intensity (D114), which when calculated based on the first expert’s score are 0.204, 0.092, and 0.646, respectively.

Due to space constraints, the individual weights of the experts for the other hierarchical single sorting work process and other indicators are not presented.

4.2.4. Maximum Eigenvalues of Judgement Matrices and Consistency Tests

Expert scores may also be inconsistent due to subjectivity and limitations, for example, there may be situations where X is more important than Y, Y is more important than Z, and Z is more important than X. In the above judgment matrix, three matrices have a 2 × 2 structure, and there is no consistency problem. The C11–D11 matrix involves four variables, and there may be inconsistencies, so it is necessary to carry out consistency testing. Now, we present the process with the first expert’s judgement matrix C11–D11₁ as an example.

The first step is to calculate the maximum eigenvalue of the matrix:

$λ_{1} = \frac{1}{4} \sum_{i = 1}^{4} \frac{{(A W)}_{i 1}}{w_{i 1}}$

(4)

${(A W)}_{1} = [\begin{matrix} 1 & 3 & 4 & 1 / 5 \\ 1 / 3 & 1 & 2 & 1 / 7 \\ 1 / 4 & 1 / 2 & 1 & 1 / 8 \\ 5 & 7 & 8 & 1 \end{matrix}] \times [\begin{matrix} 0.24 \\ 0.092 \\ 0.058 \\ 0.646 \end{matrix}] = [\begin{matrix} 0.843 \\ 0.369 \\ 0.236 \\ 2.776 \end{matrix}]$

(5)

$λ_{1} = \frac{1}{4} (\frac{0.843}{0.204} + \frac{0.369}{0.092} + \frac{0.236}{0.058} + \frac{2.776}{0.646}) = 4.121$

(6)

The second step is to calculate the consistency indicator:

${C I}_{1} = \frac{λ_{1} - n}{n - 1} = \frac{4.121 - 4}{4 - 1} = 0.04$

(7)

In the third step, calculate the consistency ratio CR and make a judgment:

${C R}_{1} = \frac{{C I}_{1}}{{R I}_{n}} = \frac{0.04}{0.9} = 0.045$

(8)

RI is the average random consistency indicator, and its value can be obtained by looking up the table. The calculation shows that CR₁ < 0.1, which shows that the C11–D11 matrix of the first expert can pass the consistency test. However, the consistency tests of the remaining 17 C11–D11 matrices using the same method found that two questionnaires could not be passed. C12–D12 of the 16 questionnaires that passed the C11–D11 matrix consistency test were also found to pass the consistency test.

4.2.5. Determine the Average Weight of Experts for the Indicator

We finally adopted the data from the 16 valid expert questionnaires that passed the consistency test and calculated the individual weights of the experts for each indicator according to the single-level ranking method mentioned above, and then averaged the individual weights of the experts as the weights of each indicator, for example, the weights of the indicator of the intensity of industrial investment were calculated using the formula:

$w_{D 111} = \frac{1}{16} \sum_{k = 1}^{16} w_{k, D 111}$

(9)

After calculation, the weights of the indicators at each level of the evaluation for city– industry integration in industrial parks are shown in Table 3.

As for the weights of the third-level indicators, among the sub-indicators of the industrial land efficiency indicator, the weight of the industrial output intensity indicator is as high as 0.508, which is nearly 1.5 times that of the sum of the other three indicators, reflecting that experts attach importance to the result-oriented evaluation of industrial land efficiency. For service sector land efficiency, the weight of the output intensity indicator reaches 0.512, which also shows a similar result. However, compared with the weight gap between the employment density and the industrial output intensity in industrial areas (0.325), the weight gap between the employment density indicator of the service area and the output intensity of the service area (0.202) is significantly smaller, reflecting the experts’ emphasis on the role of the service sector in absorbing employment.

As for the weights of the secondary-level indicators, among the sub-indicators for land–industry integration, the weight of the industrial land efficiency indicator reaches 0.597, about 1.5 times that of the service sector land efficiency, indicating that experts pay more attention to avoiding the blind “spreading out“ in industrial areas and improving the efficiency of industrial land in promoting city–industry integration.

Among the sub-indicators for residence–industry integration (integration degree of production functional area and life service functional area), the indicator weight of the matching degree between residence and environment is 0.556, and the indicator weight of the rail transit support is 0.444. The former is moderately higher than the latter, reflecting that experts pay a little more attention to promoting residence–industry integration through sufficient residential space support, and give priority to the promotion of job–residence balance. On the other hand, there is little difference between the weight of the supportive rail transit facilities and the weight of the match degree between residence and environment, indicating that experts are also aware that in the case of heavy industrial pollution in the industrial park, it is appropriate to lay out residential space and production service space in the parent city far away from the industrial park, and reduce the transportation cost between the industrial park and the parent city through rail transit, and then promote the integration of industrial parks and the parent city.

4.3. Evaluation Results of City–Industry Integration in Sample Industrial Parks

In terms of the evaluation of land–industry integration, according to the weights of the evaluation indicators for city–industry integration in industrial parks (see Table 3), the formula for the weighted score of industrial land efficiency in ETDZs is as follows:

The weighted score of industrial land efficiency = 0.211 × standardized value of industrial investment intensity + 0.183 × standardized value of industrial employment density + 0.098 × standardized density of invention patents on industrial land + 0.508 × standardized value of industrial output intensity

(10)

The formula for the weighted score of service land use efficiency is as follow:

Weighted score of service land use efficiency = 0.512 × standardized output intensity value of service sector + 0.178 × standardized density of invention patents on services land + 0.31 × standardized employment density value of service sector.

(11)

The formula for land–industry integration is as follow:

Degree of land–industry integration = 0.597 × industrial land use efficiency score + 0.403 × service sector land use efficiency score.

(12)

The calculation results and rankings based on the formulas (see Table 4) show the differences between the service land use efficiency and industrial land use efficiency of industrial parks, further reflecting the differences in the degree of integration of their origins, and to a certain extent also reflecting the level of their industrial technology and value-added rates. Among them, the Beijing ETDZ is significantly ahead in terms of industrial land efficiency, service land efficiency, and land–industry integration, which implies that it has entered the stage of a high-tech, green, developed economy. The Chengdu ETDZ is also affected by the low land efficiency of its service sector (ranking 14th) and only ranks seventh in the degree of land–industry integration. The Jiashan ETDZ, on the contrary, has a high weighted score for its service land use efficiency, ranking second, but a low score for its industrial land use efficiency, and its degree of land–industry integration ranks fourth.

For the evaluation of the residence–industry integration degree (the integration degree of the production functional area and life service functional area), according to the weights of the evaluation indicators of city–industry integration in industrial parks (refer to Table 3), there is “industrial park’s residence–industry integration degree = 0.556 × match degree between residence and environment + 0.444 × rail transit supporting score”, of which the match score between residence and environment is as described above, and is assessed based on the negative correlation between the residential area score and the Air Quality Index (AQI) of the ETDZ.

The calculated results and rankings (see Table 5) present the differences between the residence–environment matching degrees and the differences between rail transportation support in the different industrial parks, further reflecting the differences in the degree of residence–industry integration. Among them, the degree of residence–industry integration of the Chengdu ETDZ reaches 0.978 points, ranking first. Although the match degrees between the residences and environment of the Hai’an and Lianyungang ETDZs are slightly higher than that of the Chengdu ETDZ, there is a big difference between them and Chengdu in supportive rail transit, and the weighted score of residence–industry integration is significantly behind Chengdu, ranking fourth and sixth, respectively. In terms of supportive rail transit, the Beijing ETDZ is slightly higher than the Chengdu ETDZ, but its residential supporting spaces are insufficient, and the match score between residence and environment lags far behind the Chengdu ETDZ, ranking seventh in the weighted score of residence–industry integration. The match score between residence and environment in the Wuhan and Tianjin ETDZs is in the middle, but the support of rail transit stations is relatively high, and the integration score of residence–industry integration ranks second and third, respectively. The Linyi, Ningguo, Longyan, and Ningbo Daxie ETDZs rank last in the evaluation of residence–industry integration because of the low match degree between residence and environment and the absence of urban rail transit stations.

This is based on the evaluation of land–industry integration degree and residence–industry integration degree, and the weights determined by the Analytic Hierarchy Process. According to the formula “city–industry integration degree in ETDZs = 0.417 × degree of land–industry integration + 0.583 × degree of residence–industry integration”, the final weighted score is given for the city–industry integration degree of each sample economic and technological development zone (see Figure 7). Although the residence–industry integration degree of the Beijing ETDZ is only at the intermediate to high level, the degree of land–industry integration is far ahead, thus it ranks first in the score of city–industry integration. Chengdu ETDZ ranks second in terms of city–industry integration. Although its degree of land–industry integration lags behind Tianjin, Quanzhou, Jiashan, Rugao, and other ETDZs, the degree of lag is not large, but it is significantly ahead in residence–industry integration degree, and the residence–industry integration degree has a higher weight, making its city–industry integration score higher than that of the above-mentioned ETDZs. For example, compared with the Tianjin ETDZ, which ranks third in terms of city–industry integration, the degree of land–industry integration in the Chengdu ETDZ is about 0.18 points lower, but the degree of industry–residence integration is about 0.41 points higher, making the degree of city–industry integration about 0.17 points higher. The Tianjin ETDZ ranks third in the degree of city–industry integration because the land–industry integration ranks second and the residence–industry integration ranks third.

4.4. Evaluation Verification Based on Entropy Weight Method

In the above section, we evaluated city–industry integration using an analytic hierarchy process. Generally speaking, the analytic hierarchy process can synthesize expert opinions to obtain the indicator weights required to meet the objective practical work, but there may also be the subjectivity of expert judgment. To this end, we conduct another evaluation using the entropy weight method with strong objectivity to verify the reasonableness of the above evaluation results. The concept of entropy was proposed by the German physicist Clausius in 1850, and Shannon used the concept of information entropy in 1948 to express the uncertainty of signals. In information theory, entropy can measure the amount of effective information carried in a set of data; the greater the fluctuation of the data, the smaller the entropy value, the greater the amount of effective information, and vice versa.

After the data are standardized, we use the entropy weight method to calculate the weights of the indicators at each level in a bottom-up order, for example, firstly, at the level of third-level indicators, we determine the weights of industrial investment intensity, industrial employment density, density of invention patents on industrial land, and industrial output intensity in evaluating industrial land efficiency, determining the weight of the service sector output intensity, density of invention patents on services land, and service sector employment density in evaluating the service land efficiency, and then we calculate the values of two second-level indicators: the industrial land efficiency and the service sector land efficiency. Then, we determine the weights of these two indicators in evaluating the land–industry integration using the entropy weight method and calculate the values of the land–industry integration, and we finally determine the weights of the land–industry integration and the residence–industry integration and calculate the degree of city–industry integration. In the following section, the process of determining the weights using the entropy weight method is presented through the three indicators, industrial investment intensity, industrial employment density, and industrial output intensity.

From judgment matrix to normalization matrix P = (p_ij)_23×4

The evaluation objects include 23 sample development zones, and the evaluation indicators are industrial investment intensity, industrial employment density, and industrial output intensity, with the indicator units are billion yuan/square kilometer, per person/square kilometer, and billion yuan/square kilometer, respectively. After dimensionless processing using the range standardization method, the judgment matrix R is formed, and then the matrix P is normalized.

$R = [\begin{matrix} 0.147 & 0.329 & 0.166 & 0.615 \\ 1 & 1 & 1 & 1 \\ \dots & \dots & \dots & \dots \\ 0.06 & 0.126 & 0.01 & 0.522 \end{matrix}] \overset{p_{i j} = \frac{r_{i j}}{\sum_{i = 1}^{23} r_{i j}}}{\to} P = [\begin{matrix} 0.056 & 0.079 & 0.056 & 0.066 \\ 0.384 & 0.214 & 0.334 & 0.107 \\ \dots & \dots & \dots & \dots \\ 0.023 & 0.030 & 0.023 & 0.056 \end{matrix}]$

(13)

2.: Calculate the entropy matrix E from the P matrix

$P \overset{e_{j} = - \frac{1}{l n 23} \sum_{i = 1}^{23} p_{i j} {l n p}_{i j}}{\to} E = [\begin{matrix} 0.799 & 0.841 & 0.788 & 0.929 \end{matrix}]$

(14)

3.: Calculate the entropy weight matrix W

$E \overset{w_{j} = \frac{1 - e_{j}}{\sum_{j = 1}^{4} (1 - e_{j})}}{\to} W = [\begin{matrix} 0.313 & 0.248 & 0.329 & 0.111 \end{matrix}]$

(15)

For the other indicators, we also calculated their entropy weights using the above steps, see Table 6.

The weights of the indicators calculated using the entropy weighting method and the hierarchical analysis method differ considerably, but there are important similarities between the evaluation results of the two methods. Compared with the evaluation score of city–industry integration in industrial parks based on the analytic hierarchy process, the entropy weight method has a higher degree of dispersion in the evaluation scores (see Figure 8). This is because the entropy weight method, like the factor analysis method, principal component method, and other methods, gives greater weight to the indicators with stronger discrete observations, which will create a Matthew effect—the indicators that originally had a large impact on the evaluation score are doubled to expand their influence, and vice versa. However, the ranking of the city–industry integration degree of each industrial park based on the indicator entropy weight is similar to the above evaluation results (see Figure 7). The Beijing, Chengdu, and Tianjin ETDZs are also in the top three, the Ningbo Daxie and Linyi ETDZs are in the last three, and the Longyan ETDZ, which was the second-last ETDZ to be evaluated using the analytic hierarchy process, is also in the fifth-last lagging position for the entropy weight evaluation results. For the evaluation objects with a middling degree of city–industry integration, the evaluation results of the two methods are slightly more different, but the evaluation results of the two methods are highly similar in the identification of the forefront and last place ETDZs, and the entropy weight method verifies the rationality of the above analytic hierarchy process.

5. Discussion

5.1. The Enlightenment for Practice from the Benchmark Industrial Park of City–Industry Integration: From the Chengdu Model to the Beijing Model

The evaluation of city–industry integration found that the city–industry integration models and experiences of the Beijing ETDZ and Chengdu ETDZ are worthy of reference. We can refer to them as the Beijing model and Chengdu model of city–industry integration in industrial parks. The former represents a city–industry integration model for industrial parks that is based on mature green manufacturing systems used in developed economies. The latter represents the city–industry integration mode in industrial parks based on a semi-green manufacturing system used in the middle to late stages of industrialization.

The typical features of the city–industry integration model of the Beijing ETDZ mainly include (1) green development of high-tech and high value-added industries. A mature green manufacturing system, including clean production, comprehensive utilization of resources and efficient use of energy, which are fundamental prerequisites, followed by the promotion of multi-story factories to improve the floor area ratio of industrial areas to achieve efficient use of industrial land and a fresh environment—the air monitoring point located in the ETDZ shows that the air quality throughout the year is generally high quality (AQI = 46 in 2021), and is significantly better than the average air quality of the whole city (AQI = 52.58 in 2021). (2) Coordinated development of the manufacturing industry and service sector. Industrial development cannot be separated from support for logistics, research and development design, supply chain management, marketing, professional technical consulting, and other services industries. The Beijing ETDZ has laid out a large area of the service sector, which is coordinated and integrated with the green industry in the development zone. They complement each other to achieve a high level of employment density and output intensity. (3) The proportion of residential space is relatively insufficient, but the absolute number is high. Due to the high employment density in ETDZs, the per capita residential area is relatively low, and the air environment is excellent. It is more obvious that the per capita residential area layout is not enough, however, if only in terms of the proportion of residential areas in the built-up area. Still, the Beijing ETDZ is at the forefront, indicating that it has made efforts and achieved success in providing employees in the ETDZ with nearby living accommodations and promoting a balance between jobs and residences within the limits of what is possible. (4) Rail transit is convenient and comfortable. The Beijing metro Yizhuang line forms a “Z” shape within the ETDZ, providing more subway stations throughout the zone, and the Yizhuang T1 line tram runs through the Beijing ETDZ in a southwest to northeast direction. Convenient and comfortable rail transportation is conducive to reducing the time spent in traffic both inside and outside the ETDZ and is conducive to reducing people’s traffic burden. It is very important for enterprises within the ETDZ to obtain high-level productive services outside the area, so that workers in the area can access more diverse life services outside the area, and for workers outside the area to commute to. These main characteristics are summarized in Table 7.

In short, the Beijing model of city–industry integration in industrial parks takes the green development of high-tech and high value-added industries as a key prerequisite. Based on this condition, a large number of residential areas, research and development design, and other modern service areas can be located near ETDZ. Convenient and comfortable rail transit is an important condition to promote the close connection between the industrial park and other areas of the mother city. It is also a compensatory condition for the lack of per capita residential area layout.

The premise of the Beijing model is that high-tech and high value-added green industries are mainly laid out inside the ETDZ, which represents a city–industry integration mode for advanced industrial parks in developed economies. Ordinary industrial parks in developing countries need to continue to work hard to continuously improve the technology and green development level of industrial enterprises and reach the post-industrial level, so that it is possible to achieve this high-level model of city–industry integration in industrial parks. For industrial parks in developed countries or regions, the aspects that need to be learned from the Beijing model in promoting the city–industry integration include ① improving the green manufacturing system and minimizing negative externalities; ② planning and laying out sufficient modern productive services to complement industrial development; and ③ appropriate layout of residential functional areas. For industrial parks in developed countries or regions, the aspects that need to be learned from the Beijing model in promoting the integration of industry and cities include ① a sound green manufacturing system to control the negative externalities at the lowest level; ② the planning and layout of sufficient modern productive services to complement industrial development; ③ appropriate layout of the residential functional area and living services under the premise of maintaining the production function of the industrial park; and ④ construction of sufficient high-level transportation facilities such as subways to keep the industrial park in close contact with the mother city. The Chengdu ETDZ represents the city–industry integration mode in industrial parks based on a semi-green manufacturing system in the medium to late stage of industrialization, with the following characteristics: (1) the industrial development has reached the advanced manufacturing level; the leading industries are automobile and construction machinery manufacturing, and the green manufacturing system including clean production, comprehensive utilization of resources, and efficient utilization of energy is in its growth stage, and has not yet reached a mature leading level. If that is the case, then the industrial zone still has a medium to low level of negative external impact on the surrounding area. (2) Due to the negative externality of the industrial park to a certain extent, the layout of the service sector, especially the high-level modern service sector, is insufficient in the economic development zone, and the development level of the modern service sector significantly lags behind the manufacturing industry. The added value of the service sector is only 1/3 of the added value of the manufacturing industry, and it mainly relies on the mother city to provide modern productive services. The land use efficiency of the service sector area in the industrial park is not high—slightly below the medium level. (3) In view of the medium to low level negative externality of the industrial area, the proportion of the residential area to the per capita residential supporting area is at a medium to high level, taking into account the need for pollution avoidance and the need for workers to live nearby, and making a compromise between the two sides of the contradiction. (4) The convenient and comfortable subway connects the industrial park and the mother city (the core area of Chengdu city and the core area of Longquanyi district), so as to provide high-quality transportation guaranteeing that the industrial park can obtain productive services from the mother city, that the industrial park employees can commute, and that the industrial park employees living in the industrial park and their families can obtain life services in the mother city. These main characteristics are summarized in Table 7.

Most industrial parks in China and other developing countries are in the growth stage, and the level of green manufacturing is not high enough. The aspects that need to be learned from the Chengdu model in the promotion of industry–city integration include ① strengthening the land use management and trying to improve the level of industrial technology, and then improving the efficiency of the industrial land use to whatever extent is possible; ② matching the scale of the residential area according to the principle of negative correlation with the negative externalities of the industrial parks; and ③ constructing enough subways or other rail transportation facilities to keep the industrial park in close contact with the mother city. For industrial parks with insufficient local financial resources or a lack of population scale effects, urban expressways and other high level transportation methods can be used as alternatives.

With the further improvement in the level of industrialization, the industrial parks that draw on the Chengdu model to promote city–industry integration can enter a mature period. Thus, the green manufacturing system gradually matures and then the large-scale service sector and residential areas can be arranged nearby in the industrial park so that the city–industry integration model of the industrial park can be upgraded to the Beijing model.

5.2. Academic Contributions

Obviously, the Beijing model we mentioned here is also a benchmark for the industrial park city–industry integration according to the evaluation criteria of other research teams in academia, because the Beijing ETDZ has realized a high degree of spatial integration between production functional areas and living functional areas, and it and the surrounding urban areas have also realized mutual support in production and living.

On the other hand, as we mentioned in the review, established studies on industrial parks have attached great importance to the issue of the environmental impact of industrial parks on neighboring regions. In the evaluation study of industry–city integration, Wang Xia and Wang Yanhong et al. (2014) also found that some national high-tech zones do not have a high enough comprehensive utilization rate for industrial solid waste, and that the park environment is not sufficiently accommodating to new entrants [12]; Jia Jing and Bai Shanshan et al. (2019) also regarded air quality as an important factor affecting the integration of high-tech zones into industries and cities [15]. In such cases, the high spatial proximity or even mixed layout of industrial production functional areas and living functional areas in industrial parks will lead to industry–city conflicts. Our evaluation study finds that the Chengdu ETDZ attaches great importance to the matching degree between residential support and comprehensive air quality, keeps a moderate distance between a part of the living functional areas and the industrial production functional areas in the spatial layout, and realizes convenient commuting for the workers in the development zone through rail transportation, while also achieving the top degree of industry–city integration. Our research can expand the understanding of industry–city integration and we have found that the Chengdu model sets an example for other industrial parks to promote industry–city integration by correctly handling negative externalities.

Furthermore, most of the existing studies on industry–city integration focus on examining the consistency of industrialization and urbanization in prefectural or provincial areas (including the entire area of urban districts, towns, and villages under the jurisdiction of that prefecture), such as Lu Gan and Lihong Wei et al. (2022) [30] and Xinzhi Zhang and Yufei Zeng et al. (2020) [51]. From the perspective of urban areas, there are relatively few academic papers that study this issue at the scope of a certain segment of the city. Due to poor data availability, there is even less research on the evaluation of city–industry integration in industrial parks. This paper has evaluated city–industry integration in industrial parks based on the data of 23 sample industrial parks and has summarized the Beijing and Chengdu models, which helps make up for the shortcomings.

6. Conclusions

The concept of city–industry integration is proposed for the construction of new urban areas such as industrial parks, and in this urban downtown perspective, city–industry integration should be understood as a coordinated, rationally laid out, and mutually supportive relationship within and between production and service function areas centered around the urban population. Its features include intensive and efficient production function areas and intensive and efficient service function areas. Additionally, the spatial layout of production function areas and service function areas should comply with the requirements of pollution avoidance and there should be a reduction in the transportation cost between the two areas through the adoption of rail transit and other methods of transportation.

The evaluation of city–industry integration in industrial parks can be carried out from two aspects: “land–industry integration” and “residence–industry integration”. Land–industry integration in industrial parks refers to the intensive use of land for secondary and tertiary industries, and the first-level indicator of land–industry integration includes two second-level indicators, industrial land efficiency and service land efficiency, of which the industrial land efficiency indicator includes three third-level indicators, industrial investment intensity, industrial employment density, and industrial output intensity. The service land efficiency also includes two third-level indicators, service sector employment density and service sector output intensity. Residence–industry integration in industrial parks refers to the coordination of industrial production function areas and residential areas, including two secondary indicators: matching degree of residential environment and a supportive rail transport. Among them, the matching of residential areas to the environment means that the scale of residential areas in industrial parks should match with the comprehensive air quality of the industrial park, and the better the comprehensive air quality of the industrial park, the larger the residential area should be in the industrial park, and vice versa. Rail transit support is the number of rail transit stations per square kilometer within the industrial park.

We collected data from 23 national sample industrial parks through planning maps, geo-fencing, enterprise credit information databases, etc., and used the analytic hierarchy process to design expert questionnaires. We also calculated indicator weights according to experts’ evaluation information on the relative importance of the indicators. The evaluation found that the degree of city–industry integration in the Beijing, Chengdu, and Tianjin ETDZs ranked as the top three among the sample industrial parks, and the same conclusion was obtained by using the entropy weight method.

An analysis of the benchmarks for city–industry integration, the Beijing and Chengdu ETDZs, shows that the Beijing model represents the city–industry integration model of high-level advanced industrial parks in developed economies, with the green development of high-technology and high-value-added industries as the key prerequisites. Modern service areas such as large residential areas and research and development and design can be laid out close to the ETDZ based on this condition, and convenient and comfortable rail transit is an important condition to promote the close connection between ETDZs and other areas of the mother city, and it is also a compensatory condition for the insufficient layout of per capita residential areas. The Chengdu model represents the city–industry integration model of an industrial park in its growth period, with an immature green manufacturing system, and there are certain negative externalities to the surrounding environment. Negatively correlated with this, the scale of the service sector and residential area layout in the development zone is moderate, and there are many service needs and residential needs that cannot be met by the internal supporting facilities of the development zone, but they can get support from the mother city through convenient and comfortable subway transportation.

Most of the industrial parks in China and other developing countries are in the growth stage, their green manufacturing levels are not high enough, and the city–industry integration model of the Chengdu ETDZ is worth learning from. If their mother cities have the ability to build subways and other rail transportation like there is in Chengdu, then it is possible to learn from the Chengdu model in a comprehensive way; for cities with tight financial resources, the Chengdu ETDZ is still worth learning from in terms of the practice of appropriately separating the layout of industrial production functional zones and a part of the living functional zones according to the degree of negative externalities. Although this does not guarantee that it will lead in the degree of city–industry integration, at least it can help avoid city–industry conflicts and maximize the degree of city–industry integration as much as possible.

With further improvements in the level of industrialization, the industrial park can enter a mature period, the green manufacturing system can gradually improve, and sufficiently large-scale service industries and residential areas can be laid out in close proximity within the industrial park. When these conditions are met, the industrial park’s industry–city integration model can be upgraded from the Chengdu model to the Beijing model.

Author Contributions

Conceptualization, M.X.; methodology, M.X.; validation, M.X., Y.L. and D.L.; investigation, M.X., Y.L. and D.L; resources, M.X.; data curation, Y.L. and D.L.; writing—original draft preparation, M.X.; writing—review and editing, M.X., Y.L. and D.L.; visualization, Y.L.; project administration, M.X. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by National Social Science Foundation of China (grant no. 17BJY048).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are contained within the article.

Acknowledgments

Luyao Liu, Wen Li, Xuechun Wang, Xinggui Li, Zixun Mo, Zhekun Wang, and Xiangyun Wu played an important role in the data collection and processing of this paper, and we would like to express our gratitude!

Conflicts of Interest

The authors declare no conflicts of interest in this study.

Appendix A

Table A1. Beijing ETDZ coordinates to pick up part of the sample table.

Name of ETDZ	Vertex Number	Longitude	Latitude
Beijing ETDZ	2\1	116.593694686889	39.7775400221545
Beijing ETDZ	2\2	116.566790061568	39.8058337127532
Beijing ETDZ	2\3	116.563310623168	39.8012829059164
Beijing ETDZ	2\4	116.552753448486	39.7962052557995
Beijing ETDZ	2\5	116.532583236694	39.8201394667626
Beijing ETDZ	2\6	116.512670516967	39.8113711500501
Beijing ETDZ	2\7	116.502628326416	39.8184254477015
Beijing ETDZ	2\8	116.476020812988	39.8052392565111
Beijing ETDZ	2\9	116.478853225708	39.8024698349971
Beijing ETDZ	2\10	116.470527648925	39.7974582172923
Beijing ETDZ	2\12	116.492757797241	39.7848616184177
Beijing ETDZ	2\13	116.483144760131	39.7802444862761
Beijing ETDZ	2\14	116.495332717895	39.7737140133473
Beijing ETDZ	2\15	116.493530273437	39.7645439197561
Beijing ETDZ	2\16	116.481084823608	39.7655995323052
Beijing ETDZ	2\17	116.486663818359	39.7322739560859
Beijing ETDZ	2\26	116.503314971923	39.7116766334180
Beijing ETDZ	2\27	116.532339864756	39.7147385924770
Beijing ETDZ	2\29	116.530622866242	39.7422610182087
Beijing ETDZ	2\30	116.547775268554	39.7675787622142
Beijing ETDZ	2\32	116.575155258178	39.7733182071942

Appendix B

Questionnaire on the weight of industrial park city–industry integration evaluation index.

Dear Experts,

Greetings! Considering your profound expertise in the field of urban economy and industrial economy, we cordially invite you to kindly complete this questionnaire and provide valuable insights on the significance of evaluation indicators for industrial park city–industry integration. Your opinions serve as a crucial foundation for us to determine the weightage of evaluation indicators in our research on key factors and influencing mechanisms of industrial park city–industry integration. We sincerely appreciate your assistance and support towards this research endeavor.

Table A2. Numerical scaling of the relative importance of the two indicators.

Digital Scale	Implication
1	Equally important
3	One factor is slightly more important than the other
5	One factor is significantly more important than the other
7	One factor is more strongly important than the other
9	One factor is extremely more important than the other
2, 4, 6, 8	The median of the two adjacent judgments above

*1. In the indicator of industrial land efficiency, when comparing investment intensity (A, the meaning of the indicator is showed in Table A3, the same below) with employment density (B), what you think is no less important than the other indicator is ( ) (please fill in A or B, the same below), and its importance is ( ) (please fill in the figures according to Table 1, the same below).

*2. In the indicator of industrial land efficiency, comparing investment intensity (A) with industrial output per mu (B), you think that ( ) is no less important than the other indicator, and its importance is ( ).

*3. In the indicator of industrial land efficiency, comparing employment density (A) with industrial output (B), you think that ( ) is no less important than the other indicator, and its importance is ( ).

*4. In the indicator of industrial land efficiency, comparing investment intensity (A) with the density of authorized invention patents on industrial land (B), you think that ( ) is no less important than the other indicator, and its importance is ( ).

*5. In the indicator of industrial land efficiency, comparing employment density (A) with the density of authorized invention patents on industrial land (B), you think that ( ) is no less important than the other indicator, and its importance is ( ).

*6. In the indicator of industrial land efficiency, comparing Industrial output intensity (A) with the density of authorized invention patents on industrial land (B), you think that ( ) is no less important than the other indicator, and its importance is ( ).

*7. In the indicator of land use efficiency in the service sector of the industrial park, when comparing the output intensity (A) of the service sector with the employment density (B) of the service sector, you think ( ) is no less important than the other indicator, and its importance is ( ).

*8. In the indicator of land use efficiency in the service sector of the industrial park, when comparing the output intensity of the service sector (A) with the density of authorized invention patents on services land (B), you think ( ) is no less important than the other indicator, and its importance is ( ).

*9. In the indicator of land use efficiency in the service sector of the industrial park, when comparing employment density of the service sector (A) with the density of authorized invention patents on services land (B), you think ( ) is no less important than the other indicator, and its importance is ( ).

*10. In the indicator of intensity of production functional zones, comparing industrial land efficiency (A) with service land efficiency (B), you think that ( ) is no less important than the other indicator, and its importance is ( ).

*11. In the indicator of the integration degree of production and service functional zones, comparing the match score between residence and environment (A) with that of transportation facilities (B), you think that ( ) is no less important than the other indicator, and its importance is ( ).

*12. In terms of measuring the city–industry integration in industrial parks, comparing the intensity of production functional zones (land–industry integration) (A) with the integration of production and life functional zones (residence–industry integration) (B), you think ( ) is no less important than the other indicator, and its importance is ( ).

Appendix to the questionnaire: Explanation of the evaluation index system of industrial park city–industry integration.

Table A3. Indicator system and related formulas for the evaluation of industrial park city–industry integration development (A1).

First-Level Indicator (Intermediate Layer Element)	Second-Level Indicator (Intermediate Layer Element)	Three-Level Indicator (Factor Layer)	Index Calculation Formula
Intensive degree of production function areas (Land–industry integration)	Industrial land efficiency	Investment intensity	0.5 × Registered capital of unit industrial land + 0.5 × paid-in capital of unit industrial land
		Employment density	Number of people paying social security in industrial enterprises/Industrial land area
		Density of invention patents on industrial land	Authorized patents of inventions for industrial enterprises /industrial land area
		Industrial output intensity	Industrial added value/industrial land area
	Service industry land efficiency	Service output intensity	Added value of service sector/land use of service sector(built-up area—factory area—green space and water area—residential area)
		Density of invention patents on services land	Authorized patents of inventions for services enterprises/service sector land area
		Service employment density	Number of people paying social security in services enterprises/services land area
Integration degree of production functional area and life service functional area (Residence–industry integration)	Match degree between residence and environment	-	match degree between residence and environment calculation formula: 1 − \|z_r + z_a − 1\|. z_r refers to industrial park residential area support, positive indicator. z_a refers to the composite air quality index of the industrial park, inverse indicator.
	Supportive Rail transit facilities	-	Standardization of rail traffic numbers

References

Minchin, T.J. It knocked this city to its knees’: The closure of Pillowtex Mills in Kannapolis, North Carolina and the decline of the US textile industry. Labor Hist. 2009, 50, 287–311. [Google Scholar] [CrossRef]
Pallagst, K.; Fleschurz, R.; Nothof, S.; Uemura, T. PlanShrinking2-Trajectories of Planning Cultures in Shrinking Cities: The Cases Cleveland/USA, Bochum/Germany, and Nagasaki/Japan. IPS Working Paper. 2018. Available online: https://www.ru.uni-kl.de/fileadmin/intplan/Publikationen/Working_Paper_Series/IPS_WP_1_2018_-_PlanShrinking.pdf (accessed on 23 May 2024).
Schackmar, J.; Fleschurz, R.; Pallagst, K. The Role of Substitute Industries for Revitalizing Shrinking Cities. Sustainability 2021, 13, 9250. [Google Scholar] [CrossRef]
Rosenfeld, M.T.; Hornych, C. Could cities in de-industrialized regions become hot spots for attracting cultural businesses? The case of media industry in Halle an der Saale (Germany). Eur. Plan. Stud. 2010, 18, 371–384. [Google Scholar] [CrossRef]
Liu, Z.; Adams, M.; Cote, R.P.; Geng, Y.; Li, Y. Comparative study on the pathways of industrial parks towards sustainable development between China and Canada. Resour. Conserv. Recycl. 2018, 128, 417–425. [Google Scholar] [CrossRef]
Hudalah, D.; Nurrahma, V.; Sofhani, T.F.; Salim, W.A. Connecting fragmented enclaves through network? Managing industrial parks in the Jakarta-Bandung Urban Corridor. Cities 2019, 88, 1–9. [Google Scholar] [CrossRef]
Maynard, N.J.; Raj Kanagaraj Subramanian, V.; Hua, C.Y.; Lo, S.F. Industrial symbiosis in Taiwan: Case study on Linhai industrial park. Sustainability 2020, 12, 4564. [Google Scholar] [CrossRef]
Xi, J.P. Speech at the Central Urbanization Work Conference. Available online: http://www.bjzx.gov.cn/ztzl/ztxs/xsyz/ldjh/201911/t20191108_25284.html (accessed on 6 March 2023).
Kong, X.; Yang, F. The development of City-industry Integration and the Transformation and Upgrading of Development Zones. Inq. Into Econ. Issues 2013, 5, 124–128. [Google Scholar]
Xi, J.P. Several Major Issues in the Country’s Medium- and Long-Term Economic and Social Development Strategy [EB/OL]. Available online: https://www.gov.cn/xinwen/2020-10/31/content_5556349.htm (accessed on 12 March 2023).
Li, Y.X.; Zhang, Z.Y. Study on the measurement of city-industry integration and threshold effect in western region. Stat. Decis. Mak. 2021, 37, 86–90. [Google Scholar]
Wang, X.; Wang, Y.H.; Su, L. Index Evaluation System on the Degree of Production-City Integration in High-Tech Zones in China: Based on Factor Analysis and Entropy-Based Weight. J. Sci. Sci. Manag. S. T. 2014, 35, 79–88. [Google Scholar]
Dong, W.; Gao, X.; Chen, X.; Lin, L. Industrial Park Renovation Strategy in a Poverty-Alleviated County Based on Inefficient Land Evaluation. Sustainability 2023, 15, 10345. [Google Scholar] [CrossRef]
Tang, X.H. Evaluation and Suggestions of City-industry Integration in Development Zones from the Perspective of Urban Renewal. Econ. Issues Explor. 2014, 8, 144–149. [Google Scholar]
Jia, J.; Bai, S.S.; Wang, X.F. Evaluation of City-industry Integration Measure about National Hi-tech Zones of Henan Province. Areal Res. Dev. 2019, 38, 30–34. [Google Scholar]
He, Y.; Xia, Y. Evaluation and Countermeasures of City-industry Integration in Jiangsu Province. Mod. Econ. Res. 2017, 2, 72–76. [Google Scholar]
Tang, X.H. A Study for City-industry Integration Evaluation Model of Development Zone Based on Grey Correlation Analysis. Shanghai J. Econ. 2014, 6, 85–92+102. [Google Scholar]
Shi, B.; Deng, Y. Research on the Dynamic Coupling and Coordinated Development of the Integration of the Industries and the City in the Process of Resource-Based Cities. Ecol. Econ. 2017, 33, 122–125. [Google Scholar]
Shao, Z. Industry and City Integration; Springer: Berlin/Heidelberg, Germany, 2015. [Google Scholar]
Zheng, B.; Zhu, J. Evaluation of City-industry Integration Degree of National Independent Innovation Demonstration Zone. Stat. Decis. 2016, 18, 65–68. [Google Scholar]
Kuznetsova, S.N.; Romanovskaya, E.V.; Artemyeva, M.V.; Andryashina, N.S.; Egorova, A.O. Advantages of residents of industrial parks (by the example of AVTOVAZ). In The Impact of Information on Modern Humans; Springer: Berlin/Heidelberg, Germany, 2018; pp. 502–509. [Google Scholar]
Lambert, A.J.D.; Boons, F.A. Eco-industrial parks: Stimulating sustainable development in mixed industrial parks. Technovation 2002, 22, 471–484. [Google Scholar] [CrossRef]
Nuhu, S.K.; Manan, Z.A.; Alwi, S.R.W.; Reba, M.N.M. Roles of geospatial technology in eco-industrial park site se-lection: State–of–the-art review. J. Clean. Prod. 2021, 309, 127361. [Google Scholar] [CrossRef]
Park, J.Y.; Park, Y. Housing prices and environmental hazards: The effects of industrial park openings in Korea. Appl. Econ. Lett. 2024, 1–13. [Google Scholar] [CrossRef]
Pilouk, S.; Koottatep, T. Environmental performance indicators as the key for eco-industrial parks in Thailand. J. Clean. Prod. 2017, 156, 614–623. [Google Scholar] [CrossRef]
Son, C.H.; Oh, D.; Ban, Y.U. Eco-industrial development strategies and characteristics according to the performance evaluation of eco-industrial park projects in Korea. J. Clean. Prod. 2023, 416, 137971. [Google Scholar] [CrossRef]
Valenzuela-Venegas, G.; Salgado, J.C.; Díaz-Alvarado, F.A. Sustainability indicators for the assessment of eco-industrial parks: Classification and criteria for selection. J. Clean. Prod. 2016, 133, 99–116. [Google Scholar] [CrossRef]
Bellantuono, N.; Carbonara, N.; Pontrandolfo, P. The organization of eco-industrial parks and their sustainable practices. J. Clean. Prod. 2017, 161, 362–375. [Google Scholar] [CrossRef]
Park, J.M.; Park, J.Y.; Park, H.S. A review of the National Eco-Industrial Park Development Program in Korea: Progress and achievements in the first phase, 2005–2010. J. Clean. Prod. 2016, 114, 33–44. [Google Scholar] [CrossRef]
Gan, L.; Wei, L.; Huang, S.; Lev, B.; Jiang, W. Evaluation of City–Industry Integration Development and Regional Differences under the New Urbanization: A Case Study of Sichuan. Appl. Sci. 2022, 12, 4698. [Google Scholar] [CrossRef]
Shi, Y.S.; Li, J.Q.; Li, B.; Hang, T.Y. A New Approach to Evaluate the Integrated Development of City and Industry: The Cases of Shanghai and the Kangqiao Industrial Park. Buildings 2022, 12, 1851. [Google Scholar] [CrossRef]
Hao, H.M.; Ruan, L.; Yin, J.; Zhang, L.; Long, Y. Evaluation of urban industrial service facilities support based on industrial land survey data—Jiangsu Changzhou City as an example. Geogr. Res. Dev. 2021, 40, 88–93. [Google Scholar]
Fang, L.; Liu, Z.; Jin, C. How Does the Integration of Cultural Tourism Industry Affect Rural Revitalization? The Mediating Effect of New Urbanization. Sustainability 2023, 15, 10824. [Google Scholar] [CrossRef]
Zhang, J.Q.; Shen, S.W. Evaluation of city-industry integration in the middle reaches of Yangtze River urban agglomeration. Shanghai Econ. Res. 2017, 3, 109–114. [Google Scholar]
Huang, H.; Zhang, W.X.; Cui, Y.N. Evaluation and Countermeasures of City-industry Integration in Development Zones in the Context of Transformation and Upgrading—Shanxi as an Example. Econ. Issues 2018, 11, 110–114. [Google Scholar]
An, J.; Wang, R. The Coupling and Coordination Evaluation of City–industry Integration in State-level New Area—Taking Zhoushan Islands New Area and Qingdao West Coast New Area as Examples. Resour. Dev. Mark. 2021, 37, 287–293. [Google Scholar]
Cong, H.; Zou, D.; Liu, C. Spatial-temporal pattern analysis of city-industry integration in the context of new urbanization: A case study of 285 prefecture-level cities in China. Econ. Geogr. 2017, 37, 46–55. [Google Scholar]
Wang, X.; Su, L.; Guo, B.; Li, X. Evaluation of City-industry Integration of High-Tech District Based on Factor Analysis and Cluster Analysis. Sci. Technol. Prog. Policy 2013, 30, 26–29. [Google Scholar]
Zhou, Z.J.; Zhou, G.H.; Wang, Y.B.; Xiao, J. Study on the measurement of city-industry integration in the city cluster around Changzhutan. J. Nat. Sci. Hunan Norm. Univ. 2016, 39, 8–13. [Google Scholar]
Le Tellier, M.; Berrah, L.; Audy, J.F.; Stutz, B.; Barnabé, S. A sustainability assessment model for industrial parks: A Choquet integral aggregation approach. J. Environ. Manag. 2022, 316, 115165. [Google Scholar] [CrossRef]
Wang, F. Research on the evaluation of city-industry integration development in industrial agglomerations based on combined empowerment and four-grid quadrant method. Ecol. Econ. 2014, 30, 36–41+46. [Google Scholar]
Gan, L.; Shi, H.; Hu, Y.; Lev, B.; Lan, H. Coupling coordination degree for urbanization city-industry integration level: Sichuan case. Sustain. Cities Soc. 2020, 58, 102136. [Google Scholar] [CrossRef]
Li, Y.; Cao, X.; Cui, C. System Dynamics Theory Applied to Differentiated Levels of City–Industry Integration in China. Sustainability 2023, 15, 3987. [Google Scholar] [CrossRef]
Zheng, J.Q. Exploration and Practice of Space Evaluation System Based on the Concept of City–Industry Integration: A case study of urban design of Tongxiang New Town in Xiamen. Urban. Archit. 2023, 20, 82–87+112. [Google Scholar]
Li, Y. Study on Comprehensive Evaluation of the level of City–Industry Integration in Fujian Province. J. Shanxi Inst. Econ. Manag. 2022, 30, 30–36. [Google Scholar]
Kong, X.; Tian, Y. Exploration of High-spatial-resolution Integration Evaluation Method of Industry and City. In Proceedings of the 2022 29th International Conference on Geoinformatics, Beijing, China, 15–18 August 2022; Volume 29, pp. 1–6. [Google Scholar]
Luo, D.; Xiao, J. The Urban City-industry Integration Degree Evaluation Based on Ordinal Logistic Regression. In Proceedings of the 2021 7th Annual International Conference on Network and Information Systems for Computers (ICNISC), Guiyang, China, 23–25 July 2021; Volume 7, pp. 582–588. [Google Scholar]
Zou, D.L.; Cong, H.B. Spatio-Temporal Diversity Pattern of City-industry Integration and Its Influencing Factors in China. Econ. Geogr. 2019, 39, 66–74. [Google Scholar]
Perrucci, D.V.; Aktaş, C.B.; Sorentino, J.; Akanbi, H.; Curabba, J. A review of international eco-industrial parks for implementation success in the United States. City Environ. Interact. 2022, 16, 100086. [Google Scholar] [CrossRef]
Ceglia, D.; de Abreu, M.C.S.; Da Silva Filho, J.C.L. Critical elements for eco-retrofitting a conventional industrial park: Social barriers to be overcome. J. Environ. Manag. 2017, 187, 375–383. [Google Scholar] [CrossRef] [PubMed]
Zhang, X.Z.; Zeng, Y.F.; Li, X.H. Evaluation Study on the Development of Industry City Integration Driven by Manufacturing Industry Transfer. Jiangxi Soc. Sci. 2020, 40, 105–115. [Google Scholar]

Figure 1. The connotation of city–industry integration.

Figure 2. Desirable distance between production function areas and service function areas.

Figure 3. Data collection steps.

Figure 4. Sample of economic and technological development zones.

Figure 5. The air quality composite index and supportive residential area score of the sample ETDZs.

Figure 6. Matching scores between residence and environment in ETDZs.

Figure 7. City–industry integration score of sample ETDZs by AHP. Note: See Section 3 for data sources, Section 2 of this chapter for indicator weights, and the main text of this section for the calculation process.

Figure 8. Evaluation score for industry–city integration in sample industrial parks using the entropy weight method.

Table 1. Evaluation index system of city–industry integration of industrial parks.

First-Level Indicator (Criterion Layer B)	Second-Level Indicator (Sub-Criterion Layer C)	Three-Level Indicator (Elements Layer D)	Index Calculation Formula
Land–industry integration (Coordination degree and balance between internal industries and carriers of production function zones and ser-vice function zones, B1)	Industrial land efficiency (C11)	Investment intensity (D111)	0.5 × Registered capital of unit industrial land + 0.5 × paid-in capital of unit industrial land
		Employment density (D112)	Number of people paying social security in industrial enterprises/Industrial land area (factory area)
		Density of invention patents on industrial land (D113)	Authorized patents of inventions for industrial enterprises/industrial land area
		Industrial output intensity (D114)	Industrial added value/industrial land area
	Service industry land efficiency (C12)	Output intensity of the service sector (D121)	Added value of service sector/land use of service sector, where land area of service sector = built-up area—factory area—green space and water area—residential area
		Density of invention patents on services land (D122)	Authorized patents of inventions for services enterprises/service sector land area
		Services employment density (D123)	Service employment density = Number of people paying social security in service sector enterprises/service sector land area
Residence–industry integration (Coordination and integration of production functional areas and residential service functional areas, B2)	Matching degree be-tween residence and environment (C21)	Residential area supporting scale (z_r)	The standardized value of per capita residential area × 0.5 + the standardized value of the proportion of residential area to built-up area × 0.5	match degree be-tween residence and environment calculation formula: \|z_r + z_a − 1\|
	Matching degree be-tween residence and environment (C21)	Air Quality Composite Index (z_a)	AQI standardized value of industrial park× 0.5+ standardized value of (AQI of industrial park ÷ AQI of the mother city of industrial park) × 0.5
	Rail transit supporting facilities (C22)		The range standardization of “number of rail transit stations/built-up area of ETDZs”

Note: While the per capita residential floor area is a suitable measure of living facilities, many real estate projects in industrial parks cannot be found online. Therefore, we use the residential area based on land measurements from the map to measure the residential facilities in the developed area.

Table 2. Range standardized value of basic indicators of city–industry integration in sample industrial parks.

ETDZs	Industrial Investment Intensity	Employment Density in Industrial Area	Patent Density of Industrial Inventions	Industrial Output Intensity	Service Output Intensity	Service Employment Density	Patent Density of Inventions in the Service Sector	Air Quality Composite Index	Match Degree between Residence and Environment	Rail Stations per Unit Area
Tianjin	0.147	0.329	0.166	0.615	0.743	0.249	0.099	0.519	0.287	0.44
Beijing	1.000	1.000	1.000	1.000	0.912	1.000	1.000	0.168	0.337	1.00
Nantong	0.056	0.052	0.090	0.171	0.084	0.025	0.007	0.391	0.384	0.00
Kunshan	0.024	0.059	0.132	0.554	0.405	0.056	0.043	0.493	0.679	0.000
Ningguo	0.030	0.127	0.053	0.088	0.535	0.463	0.017	0.141	0.511	0.000
Daya Bay	0.024	0.015	0.004	0.162	0.069	0.004	0.000	0.328	0.742	0.000
Kunming	0.055	0.064	0.034	0.182	0.076	0.062	0.029	0.370	1.000	0.180
Ningbo Daxie	0.066	0.064	0.128	0.300	0.167	0.034	0.003	0.417	0.001	0.000
Chengdu (Damian)	0.127	0.242	0.098	0.612	0.117	0.059	0.009	0.445	0.530	0.987
Rugao	0.095	0.174	0.136	0.866	0.331	0.011	0.008	0.644	0.299	0.000
Quanzhou	0.154	0.474	0.222	0.720	0.542	0.133	0.096	0.505	0.385	0.000
Jiashan	0.056	0.166	0.094	0.446	1.000	0.107	0.021	0.523	0.401	0.000
Zouping	0.212	0.158	0.014	0.207	0.088	0.009	0.001	0.679	0.128	0.000
Lianyungang	0.001	0.060	0.099	0.102	0.029	0.037	0.002	0.558	0.421	0.000
Hanzhong	0.019	0.030	0.005	0.580	0.229	0.033	0.000	0.677	0.574	0.000
Korla	0.053	0.145	0.000	0.629	0.000	0.003	0.000	0.839	0.058	0.000
Zhangjiagang	0.173	0.465	0.247	0.358	0.138	0.042	0.018	0.482	0.692	0.000
Linyi	0.033	0.047	0.028	0.061	0.008	0.000	0.001	0.625	0.693	0.000
Longyan	0.000	0.000	0.076	0.000	0.213	0.120	0.045	0.372	0.256	0.000
Hai’an	0.053	0.097	0.191	0.346	0.083	0.013	0.008	0.644	0.366	0.423
Wuhan	0.128	0.181	0.135	0.474	0.056	0.013	0.009	0.460	0.356	0.436
Zhengzhou	0.040	0.080	0.036	0.355	0.132	0.046	0.005	0.692	0.426	0.236
Changchun	0.060	0.126	0.010	0.522	0.304	0.009	0.002	0.417	0.870	0.539

Table 3. The weight of the evaluation index for city–industry integration in industrial parks by AHP.

First-Level Indicator (Guideline Layer B)	Weight (w)	Second-Level Indicator (Sub-Guideline Layer C)	Weight (w)	Three-Level Indicator or Definitions (Element Layer D)	Weight (w)
Land–industry integration (Coordination degree and balance between internal industries and carriers of production function zones and service function zones, B1)	0.417	Industrial land efficiency (C11)	0.597	Industrial investment intensity D111	0.211
				Industrial employment density D112	0.183
				density of invention patents on industrial land D113	0.098
				Industrial output intensity D114	0.508
		Service industry land efficiency (C12)	0.403	Service industry output intensity D121	0.512
				density of invention patents on services land D122	0.178
				Services employment density D1232	0.31
Residence–industry integration (Coordination and integration of production functional areas and residential service functional areas, B2)	0.583	Matching degree be-tween residence and environment (C21)	0.556	The degree of negative correlation between residential area size and AQI
	0.583	Rail transit supporting facilities (C22)	0.444	Rail transit station per unit area

Table 4. Evaluation of land–industry integration in sample industrial parks by AHP.

ETDZs	Land–Industry Integration Weighted Score with Ranking		Industrial Land Efficiency Weighted Score and Ranking		Weighted Score and Ranking of Service Sector Land Use Efficiency
Beijing	0.982	1	1.000	1	0.955	1
Tianjin	0.442	2	0.420	4	0.475	3
Quanzhou	0.438	3	0.507	2	0.336	5
Jiashan	0.387	4	0.278	12	0.548	2
Rugao	0.372	5	0.505	3	0.174	7
Kunshan	0.279	6	0.310	9	0.232	6
Chengdu	0.266	7	0.392	5	0.080	14
Changchun	0.244	8	0.302	11	0.159	8
Hanzhong	0.233	9	0.304	10	0.128	10
Ningguo	0.217	10	0.079	20	0.420	4
Korla	0.214	11	0.357	6	0.001	23
Zhangjiagang	0.231	12	0.328	7	0.087	12
Wuhan	0.201	13	0.314	8	0.034	20
Zhengzhou	0.157	14	0.207	14	0.083	13
Ningbo Daxie	0.153	15	0.190	15	0.096	11
Hai’an	0.153	16	0.223	13	0.048	17
Zouping	0.127	17	0.180	16	0.048	18
Kunming	0.097	18	0.119	17	0.063	15
Nantong	0.090	19	0.117	18	0.052	16
Daya Bay	0.069	20	0.090	19	0.036	19
Longyan	0.067	21	0.007	23	0.154	9
Lianyungang	0.054	22	0.073	21	0.027	21
Linyi	0.031	23	0.049	22	0.005	22

Note: See Section 3 for data sources, Section 2 of this chapter for indicator weights, and the main text of this section for the calculation process.

Table 5. Evaluation of residence–industry integration in ETDZs.

ETDZs	Weighted Score and Ranking for Residence–Industry Integration		Match Degree between Residence and Environment and Ranking		Standardized Scores and Rankings of Rail Transit Stations per Unit Area
Chengdu	0.978	1	0.971	3	0.987	2
Wuhan	0.579	2	0.694	12	0.436	5
Tianjin	0.570	3	0.677	14	0.437	4
Hai’an	0.556	4	1.000	1	0	8
Zhengzhou	0.554	5	0.809	9	0.236	6
Lianyungang	0.543	6	0.978	2	0	8
Beijing	0.529	7	0.151	22	1.000	1
Changchun	0.525	8	0.514	17	0.539	3
Rugao	0.509	9	0.916	4	0	8
Daya Bay	0.497	10	0.894	5	0	8
Jiashan	0.491	11	0.884	6	0	8
Korla	0.465	12	0.837	7	0	8
Quanzhou	0.458	13	0.824	8	0	8
Kunshan	0.397	14	0.715	10	0	8
Zhangjiagang	0.396	15	0.712	11	0	8
Zouping	0.377	16	0.678	13	0	8
Nantong	0.346	17	0.622	15	0	8
Hanzhong	0.320	18	0.577	16	0	8
Kunming	0.285	19	0.370	20	0.180	7
Linyi	0.255	20	0.459	18	0	8
Ningguo	0.226	21	0.408	19	0	8
Longyan	0.204	22	0.366	21	0	8
Ningbo Daxie	0	23	0	23	0	8

Note: See Section 3 for data sources, Section 2 of this chapter for indicator weights, and the main text of this section for the calculation process.

Table 6. Entropy weight of evaluation indicators for industrial park city–industry integration.

First-Level Indicator	Entropy Weight (w)	Second-Level Indicator	Entropy Weight (w)	Three-Level Indicator or Interpretations	Entropy Weight (w)
Land–industry integration	0.577	Industrial land efficiency	0.373	Industrial investment intensity,	0.313
				Industrial employment density	0.248
				Density of invention patents on industrial land	0.329
				Industrial output intensity	0.111
		Service industry land efficiency	0.627	Service industry output intensity	0.142
				Density of invention patents on industrial land	0.549
				Services employment density	0.309
Residence–industry integration	0.423	Matching degree be-tween residence and environment	0.1	The degree of negative correlation between residential area size and AQI
Residence–industry integration	0.423	Rail transit supporting facilities	0.9	Rail transit station per unit area

Table 7. Main characteristics of industry–city integration in the Beijing and Chengdu ETDZ.

	Industrial Technology Level	Green Manufacturing Maturity	Negative Externality	Service Sector Development	Residential Area Ratio	Residential Land Area per Capita	Density of Rail Transit Stations
Beijing	extremely high	extremely high	extremely low	high	medium to high	medium to low	high
Chengdu	medium	medium to high	medium to low	medium	medium to high	medium to high	high

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).