A Solution-Extracted System for Facilitating the Governance of Urban Problems: A Case Study of Wuhan

Abstract

Recently, rapid urbanization around the world has spawned several urban problems. Although a large amount of experience has been accumulated throughout the process of global urban problem governance, the knowledge has not been optimally utilized. Furthermore, there is a dearth of mechanisms with which to distill and employ past experiences in addressing emerging urban problems. Consequently, in this study, based on the CBR method, we establish a mechanism called the Solution-Extracted System of Urban Problem Governance (SESUPG), aiming to find solutions to the diverse array of existing urban problems from previous experience. The main steps for obtaining a suitable solution for a specific urban problem in a target city through the SESUPG are as follows: (1) Calculate the similarity to retrieve the most similar cities. (2) Extract the possible solution through similar cities. (3) Case–solution modification before solution adoption. To verify the effectiveness of the proposed mechanism, the air pollution problem in Wuhan, China, was tested to verify the effectiveness of the SESUPG as a case study. As a result, four policy recommendations were extracted by the SESUPG, and all of them proved to be effective in mitigating air pollution problems in Wuhan. The system proposed in this study can aid decision makers in the selection of strategies and solutions when addressing urbanization issues and guiding the process of mining effective experience for the promotion of urban governance levels.

1. Introduction

In 1900, only 15% of the world’s population lived in towns and cities, while only 120 years later the worldwide urbanization rate reached 56%, and it is expected to increase to 72% by 2050 [1,2]. Admittedly, the rapid urbanization process has been accompanied by rapid economic growth, which provides incentives for city development and generates diverse societal advantages [3,4]. However, the rapid expansion of cities has also triggered several new challenges from both ecological and social perspectives, including deteriorating ecological conditions [5], increasing traffic congestion [6,7], daunting housing prices [8,9], insufficient public services [10,11], fierce competition for limited-quality resources [12], a growing gap between the rich and the poor [13,14], and rising urban crime rates [15].

These emerging challenges are collectively referred to as “urban problems” [1,2,16]. Such urban problems arise as consequences of the rapid urbanization process, carrying significant side effects that demand attention. They are not only detrimental to urban sustainability but also have a profound impact on the quality of life of urban residents. For example, the air pollution caused by PM_2.5 influences more than 2600 urbanization areas in China; furthermore, it has been concluded that every 10 g/m³ increase in the annual PM_2.5 concentration will increase the all-cause mortality rate by about 3.8%. Furthermore, in the central area of London, over one million people and 40,000 motor vehicles per hour enter and exit the city center during peak hours each day. This has led to severe traffic congestion in the area, with an average vehicle speed of only 14.3 km per hour. As the rapid pace of the global urbanization process is irreversible, these urban problems will continue to pose significant challenges to urban governance worldwide, thus obstructing urban sustainability. Cities are vulnerable and impressionable when they fail to adapt to these new challenges [17]. Thus, it is imperative to take action to tackle these problems.

The emergence of urban problems has captured the attention of both scholars and city administrators. As a result, several international organizations, including the United Nations, UN-Habitat, and UNSDG have taken steps to collect excellent urban problem-solving cases from around the world to provide references for addressing urban problems [18,19,20].

It is also appreciated that previous researchers have managed to provide valuable insights into developing constructive urban planning theories to alleviate the detrimental effects of the rapid urbanization process from a macroscopic perspective. Just as the famous garden city theory proposed by Ebenezer Howard advocates integrating urban areas with the natural environment, in aiming to address the urban problem [21], the organic evacuation theory [22] and satellite city theory [23] advocate for decentralization of the center area of the city to alleviate the problems caused by population expansion in the city center. These efforts have laid the theoretical foundation for urban problem control and governance.

As cities continue to grow and grapple with uncertainties and challenges due to rapid urbanization progress, urban resilience, low-carbon cities, and smart cities—which are closely associated with urban sustainability—have become increasingly favored concepts. Urban resilience refers to the ability of an urban system to maintain or rapidly return to desired functions in the face of a disturbance, which is always related to ecological systems, climate change infrastructure systems, and emergency operations [24,25,26]. Similar to urban resilience, studies on low-carbon cities are dedicated to achieving energy efficiency and structural reform, which not only conserves energy but also alleviates environmental pollution within urban contexts [27]. Additionally, the development of smart cities drives transformative shifts in urban transportation systems, addressing issues stemming from rapid urbanization such as traffic congestion and inadequate infrastructure [28]. These concepts fundamentally aim to enhance urban sustainability and alleviate urban problems by reshaping existing urban structures to better accommodate the pace of rapid urbanization.

Furthermore, some scholars have started from an empirical analysis perspective, aiming to explore the connections and causality between urbanization procedures and various urban problems, including the environmental deterioration perspective [29], traffic conjunction perspective [30], insufficient public infrastructure [31], and increasing housing prices [32]. These studies are usually based on time-series data or panel data, combining linear or nonlinear empirical analysis models to explore the evolving patterns or analyze the correlation of urban problems with changes in urbanization rates. For example, Wang et al. [33] explored the impact of urbanization on the coupling of economic growth and environmental quality based on panel data from 134 countries from 1996 to 2015. This approach enables a better understanding of the patterns of urban ailment development and facilitates the resolution of urban problems.

Admittedly, previous works are characterized by their constructive nature and value, exemplifying the advancements in this area, while they often confine their investigation to specific issues within localized research domains. While these endeavors have yielded commendable outcomes, their generalizability remains constrained due to the inherent uniqueness of each urban setting. Blindly applying their findings without careful consideration of specific urban circumstances could lead to misapplication. In addition, the predominant focus of these inquiries is focusing on specialized problem domains, such as air pollution or traffic congestion, and the knowledge has not been systematically integrated and utilized. Consequently, the limitations of the research outcomes become manifest. Therefore, it is crucial to establish a mechanism that comprehensively integrates previous wisdom to tackle urban problems and foster urban sustainability.

In this research, we innovatively devised the Solution-Extracted System of Urban Problem Governance (SESUPG) to provide strategic solutions for urban problems. The innovation of this approach is manifested in the following aspects. (1) Each case within the SESUPG case library represents a distinctive description of urban problems, comprising urban characteristics and causative factors of urban problems. This enables the SESUPG to comprehensively consider the unique features and specific causes of urban problems in an individual city, and is thus beneficial to the strategy extraction procedure. (2) Crucially, the SESUPG does not rely on an abundance of mathematical models or formulas, and it operates as a knowledge-integrated system. The solutions within the SESUPG database have undergone practical validation, enhancing the credibility and effectiveness of governance strategies. (3) Furthermore, within the context of refining matching strategy revisions, this study introduces three innovative refining strategies: local rejection, local modification, and local consolidation. This enhances the applicability of extracted strategies for urban problem resolution. Additionally, the SESUPG supports continuous learning and improvement through the expansion of a case library, allowing for the gradual refinement of urban ailment governance strategies over time.

The remainder of this research article is structured as follows: Section 2 puts forward the theoretical basis of this study; Section 3 states the integrated framework and each step of the SESUPG; Section 4 displays the feasibility of the SESUPG and shows the concrete process with a case study of Wuhan City in central China; Section 5 presents the discussion, and Section 6 showcases the conclusions of the study.

2. Theoretical Basis

2.1. Urban Problems

There is no denying that urbanization progress has brought forth a range of problems. These include deteriorating ecological conditions, increasing traffic congestion, daunting housing prices, insufficient public services, fierce competition for limited-quality resources, and poverty. These are often referred to as “urban problems” [34]. Although the concept of urban problems lacks a unified definition in the academic literature at the current stage, there is consensus that the urban problem is a negative phenomenon that symbolizes the imbalance and disorder of rapid urbanization, leading to the inhibition of long-term city growth [35].

However, the control and governance of urban problems are complex tasks and the formation of urban problems can result from a combination of multiple factors. Taking air pollution as an example, this involves urban characteristic factors such as the urbanization rate of the city, population, geographical location and climate, and economic level. In addition, the causes of air pollution are varied, including heavy-pollution industries [36], increasing private cars [37], unregulated waste incineration [38], and adverse weather conditions [39]. Adding to the complexity, the formation of air pollution is not always in direct correspondence with the characteristics of a city or attributed to a single cause. Instead, the relationships among these factors are intertwined with one another as illustrated in Figure 1.

This intricacy makes the formulation of strategies for urban problem governance even more challenging through conventional mathematical formulas and statistical approaches. CBR performs well in handling complex problems as opposed to complex mathematical formulas for resolution formulation [40].

2.2. CBR (Case-Based Reasoning)

Case-based reasoning, first proposed by Schank [41], is a problem-solving paradigm that emphasizes utilizing past experiences to address a new problem [42,43]. Typically, a case library is created to support the entire CBR procedure, which helps to build associations between the target problem and the most similar example. Additionally, the target case can be added to the case base to enlarge the ‘think tanks’ [40]. The system has been effectively applied in a variety of sectors, including medical diagnosis [44,45], engineering project design and management [46,47], and construction management [48,49].

Generally, in addition to the preparation step, there are five primary steps in a traditional CBR process which include case representation, case retrieval, case reuse, case revising, and case retaining (See Figure 2): [50].

(1)

Case Representation (preparation step):

Case representation, although it is a preparatory step, is an essential part of a typical CBR system. In the case representation step, the gathered cases are standardized with uniform attributes and stored within the case library. The target case is also represented with the same attributes as the stored cases in the library. This is to facilitate subsequent similarity calculation and matching, as well as to enable the system to better handle and compare the characteristics among different cases. Additionally, a real-world input section that includes feedback and interpretations after implementing the solutions can be included in a case study [51,52].

(2)

Case Retrieval:

The case retrieval step is considered the most important as it identifies the cases that are most similar to the target case. Various similarity measurements are utilized to determine the degree of proximity between the target case and the stored cases. K-nearest neighbor (K-NN) and the local–global similarity technique are the two most popular methods employed in similarity measures [53,54]. The method for measuring similarity will be proposed by determining the similarity of the features based on their distinct type and structure. Then, the outcome with a high degree of similarity will be retrieved.

(3)

Case Reuse:

The process of reusing a similar case in the current stage involves two critical aspects: firstly, identifying the differences between the extracted case and the target case, and secondly, determining which elements of the extracted case can be transferred to the new case. These two components serve as the foundation for the subsequent stage [40].

(4)

Case Revise:

The process of case revise involves the conversion of potential solutions retrieved from prior cases into solutions that can be applied to the current problem context. In cases where the solutions identified in the retrieved cases do not fully meet the needs of stakeholders, the current situation must be adjusted accordingly.

(5)

Case Retain:

After successfully applying the solution to the target problem, the final target case will be added to the case library as a new case. Consequently, a CBR loop has been generated.

Figure 2. Typical CBR framework [55].

Figure 2. Typical CBR framework [55].

CBR performs well in handling complex problems and ambiguous situations because it can extract valuable information from experience and previous successful practice as opposed to complex mathematical formulas for resolution formulation [43,51,52]. CBR has the advantage of transcending the intricate interrelationships among factors and instead employs the characteristics of the target for similarity matching, thereby extracting solutions from previous experience. These solutions stem from past successful experiences and have undergone practical validation, enhancing the credibility and practicality of governance strategies. Hence, the CBR framework is well suited to urban problem governance.

In the past, scholars have applied the CBR framework to sustainable management processes. For example, Shen et al. [56] initially implemented the CBR approach for supporting decisions in the field of sustainable urbanization and building the foundation. The ExMS provides a mechanism for capturing and sharing best practices, enabling decision makers to select solutions that have proven effective in similar situations. Based on this research, Shen et al. [42] introduced the concept of urban features to the ExMS and proposed an innovative way to measure urban similarity, which can cover the deficit caused by adopting experience indiscriminately. Wang et al. [43] applied the principle of CBR to develop a Lessons Mining System (LMS) that helps government administrators make informed decisions by extracting lessons from past practices. In addition, scholars such as Tan et al. [57] and Ochoa et al. [58] investigated the content of various best practices to identify the fundamental methods, outcomes, and interrelationships among them. These studies have all employed the CBR method to achieve the reuse of past experiences to solve new problems.

Based on the CBR method, this study goes further than previous research by introducing the SESUPG, which comprehensively integrates the intrinsic characteristics of cities and the causative factors of urban problems. These two factors serve as the criteria for case retrieval, facilitating the extraction of optimal urban problem resolution strategies. By employing this approach, those challenges can be effectively addressed, thereby enhancing urban sustainability.

3. The Framework of the Solution-Extracted System of Urban Problem Governance (SESUPG)

Based on the CBR framework, in this study, we propose the SESUPG under urban problem governance scenarios. The system comprises five core components: case representation, case retrieval, case reuse, case revising, and case retaining (see Figure 3). Case representation (the preparation step of the SESUPG) lays the foundation for all subsequent work, which transforms case information into a unified structure to facilitate subsequent similarity calculation. Case retrieval, achieved through similarity calculations, matches the target case with the most suitable series of strategies for urban problem resolution. This step forms the core of the SESUPG. In the case of the reuse step, these extracted strategies are evaluated for “reuse” or “reuse with adaptations” in the target case. The case revising procedure can enhance the applicability of strategy extraction by adapting to the target case characteristics, while case retaining empowers the SESUPG with self-learning capabilities, continuously expanding and refining the system as it is used, leading to ongoing improvement. All five parts are essential components of the system. Each stage of the framework is elaborated upon in detail below.

3.1. Case Representation

The case base of the SESUPG was sourced from various governance (e.g., New York City Global Partners, Shanghai Urban Planning Program) and organizational websites (e.g., UN-Habitat, Urban Agenda Platform), as well as authoritative journals and books that provide examples of urban problem governance. To ensure relevance, these cases underwent rigorous examination and scrutiny, and significant information was then structured into a digital chart framework. The resulting SESUPG case library currently contains 326 cases. Through the case representation procedure, the stored case information can be transformed into a unified structure to facilitate subsequent similarity calculation.

The representation step primarily consists of two parts: the problem section and the solution section. The former step contains a description of the problem along with relevant city features and causes of urban problems, whereas the latter may include the solution itself [52,53].

(1)

Problem components

The problem section of this study offers a background to the case, comprising the description of urban features and the causes of urban problems.

Urban features:

Urban features comprise the identity that differentiates a city from its other counterparts [59]. Cities are complex systems comprising components, including environmental, economic, and social factors [13,60]. However, explaining the background of a city is challenging due to its multidimensionality. Many scholars have adopted various urban features or indicators to describe a city from different aspects [56,61]. Thus, in this study, we integrated previous scholarly research and combined the context of urban problem governance, extracting six related urban characteristics (see

Table 1. Urban features and their indicators/attributes.

Urban Feature	Indicator/Attribute
Urban size (C1)	Area zone (km2)
Urban scale (C2)	Resident population (10,000 people)
Urbanization stage (C3)	Urbanization rate (%)
Economic performance (C4)	GDP (billion dollars)
Topography (C5)	Hills/mountains/plains/plateaus
Climate (C6)	Monsoon/desert/rainy/frigid

) that have a close relationship with urban development and governance.

The causes of urban problems:

We applied a literature review approach to investigate the causes of each urban problem. We focused on six common urban problems (air pollution, traffic, housing, urban inequality, insufficient public services, and resource scarcity). The primary causes of urban problems were taken from studies conducted by the relevant authorities and are reported in

Table 2. The causes of common urban problems.

Category	Cause	References
Air pollution	1. High energy consumption, high pollution from traditional energy use 2. Increase in the number of motor vehicles 3. Excessive greenhouse gas emissions from sources brought about by buildings 4. With the rapid development of industry, industrial plant waste emissions 5. Unclear and confusing responsibilities for environmental governance 6. Impact of adverse climatic conditions	[61,62,63,64,65]
Traffic	1. Limited carrying capacity of roads and other infrastructure 2. Increased number of cars leads to severe traffic congestion 3. The number of parking spaces is low and parking is difficult 4. Irrational road planning and confusing management 5. Public rail transit construction lags or public rail transit operates 6. Insufficient and outdated transportation infrastructure and high maintenance costs	[6,61,66]
Housing	1. House prices make it expensive to rent 2. The impact of immigration and refugee flows 3. Unregulated housing rental market/rent control 4. Insufficient housing stock supply and housing resource scarcity 5. Absence of relevant departmental subjects 6. Inadequate legal system for related rental housing	[67,68,69]
Urban inequality	1. Excessive urban development leads to unbalanced resource allocation 2. A high unemployment rate in the urban population 3. Unequal distribution of social income and increase in the share of the population with low income 4. Lack of strong policies to support 5. The absence of government management and the lack of interdepartmental coordination mechanisms 6. The gap between urban and rural areas widening	[61,70]
Insufficient public service	1. Deviations and disorders in urban planning 2. Backward and inefficient urban governance 3. Slow construction of investment system environment hinders investment model innovation 4. Lack of proper management and maintenance of facilities 5. Lack of effective policy support	[71,72]
Resource scarcity	1. Excessive population growth and limited resource-carrying capacity 2. Lack of resources due to the geographical location of the city 3. Lack of effective management 4. High energy consumption, high pollution from traditional energy use	[73,74,75]

.

(2)

Solution components

The solution section of a case describes in detail the methods, measures, and processes used for managing urban problems. The information is presented in a textual format.

All cases saved in the library are organized uniformly. Taking a case from Los Angeles as an example, the case representation template is shown below (see

Table 3. Case representation template.

Case information:
Case number:	No. 1
City/town:	Los Angeles
Country:	America
Corresponding urban problems:	Air pollution
Urban features
Area zone (km2):	1213.8
Population (10,000 people):	747.8 (1980)
Topography:	Plain
GPD (billion dollars):	3842 (1980)
Urban rate (%):	100%
Climate:	Mediterranean climate
The causes of the problem
The traditional energy-supply-oriented industrial structure bringing high energy consumption, high pollution, and high emissions The rapid development of industrialization; industrial plant waste emissions The number of motor vehicles increasing The lack of laws and regulations to guide
Solutions
Establish a cross-administrative joint prevention and control agency to collaborate on regional air pollution problems, and establish a cross-realization air quality authority to unify the regulation of pollutant emissions from enterprises and stationary sources in the region. Formulate relatively sound laws and regulations on air pollution management; form a framework of laws and regulations at different levels such as the federal government, state governments, cross-administrative regions, local governments, and governments at all levels to formulate relevant air quality regulations and policies according to their authority and responsibilities; formulate relevant air quality regulations and policies. Give play to the prominent role of market mechanisms in environmental pollution management, attach importance to technological innovation in air pollution management, and rely on technological innovation and market forces to jointly promote pollution management. Enhance motor vehicle exhaust emission testing: mandatory installation of particulate matter filters on diesel vehicles, use of road remote sensing detection systems, truck onboard diagnostic systems, etc. for exhaust emission testing.

). A partial example of the library is illustrated in Figure 4.

3.2. Case Retrieval Mechanism

Case retrieval is the most important procedure in the SESUPG. The retrieval progress in the SESUPG includes two components: the local similarity measurement and the global similarity measurement between the target case and the stored cases (as shown in Figure 5).

3.2.1. Local Similarity Measurement of the Urban Features

The value of urban features can be categorized into two types of representations: crisp numbers, which denote features with a precise numerical value, and crisp symbols, which represent features with specific meanings [76,77,78]. The calculation methods used to quantify these two types of representations may differ.

Regarding the quantification of urban features, such as GDP, it is possible to compute their value using the following equation:

$$l{s}_{(c_0,c_j)}^j=1-|v_i^j-v_0^j|/(\overline{v_i^j}-\underset{\_}{v_i^j})$$

(1)

$l{s}_{(c_0,c_j)}^j$ signifies the distance between the target case $c_0$ and the case $c_j$. The term $v_i^j$ represents the attribute $j$ of case $i$ while $v_0^j$ refers to the attribute $j$ of the target case. The denominator of the equation incorporates the values $\overline{v_i^j}$ and $\underset{\_}{v_i^j}$, which represent the maximum and minimum values of attribute j across all cases, respectively.

The following formula is adopted for crisp symbol-type features such as topography and climate:

$$l{s}_{(c_0,c_j)}^j=\{\begin{array}{l}1if{v}_0^j=v_i^j\\ 0if{v}_0^j\ne v_i^j\end{array}$$

(2)

$l{s}_{(c_0,c_j)}^j$ signifies the distance between the target case $c_0$ and the case ${\mathrm{c}}_i$. The features can be denoted as $v_i$ and $v_0$. If the two values are the same, $l{s}_{(c_0,c_j)}^j$ is 1; otherwise, $l{s}_{(c_0,c_j)}^j$ is 0.

The local similarity of urban features between two cities can be retrieved using the following formula:

$$Si{m}_{local1}={\displaystyle \sum _{i=1}^6{\omega }_i\times l{s}_{(c_0,c_j)}^j}$$

(3)

where ${\omega }_i$ denotes the weighting of the case i. Based on the opinions of experts in relevant fields, the weighting values between six urban features were determined to be $\omega$ = (0.17; 0.16; 0.16; 0.16; 0.16; 0.17)^T. $Si{m}_{local1}$ represents the integrated similarity between the target case and the case in the library, which is the sum of the product of distance and weight of the six urban features, respectively.

3.2.2. Local Similarity Measurement of the Causes of Urban Problems

Since the description of urban problem causes is textual, the method for calculating similarity is different compared to the method for numerical urban features. In this paper, the cosine similarity method is applied to calculate the similarity between the urban problem causes in the target case and the urban problem causes in the sample case of the case library [79,80]. Cosine similarity has been widely applied in solving different text mining problems, such as text classification, text summarization, information retrieval, question answering, and so on.

In this study, the similarity calculating procedures of the urban problems causes are as follows:

Step 1: Extraction of core etiological terms for the specific urban problem.

Urban problem causes are sourced from literature and news reports and subsequently consolidated and refined. The subjects of sentences are then extracted as core terms, as they often encapsulate the fundamental causes behind urban problems. For example: “the traditional energy-supply-oriented industrial factory emission bringing high pollution: if excessive factory emission continues, the air condition will deteriorate.” In this sentence, “industrial factory emission” is mentioned in the text; thus, the expression would be considered a core term.

Step 2: List all core terms for a specific urban problem.

By following the first step, all the causative factors related to urban problem P (assuming it is air pollution) are extracted and listed from the text. The results are as follows (excessive factory emissions; vehicle exhaust emissions; unfavorable climate conditions; agricultural activities; massive utilization of fossil fuels; insufficient legislation; construction and demolition dust; and natural sources (e.g., dust storms, wildfires)).

Step 3: Generate word vector.

For each core word, a binary value of 1 is assigned if the case matches with the core word; a value of 0 is assigned if it does not. The results are then represented as a vector. The length of the vector is determined by the total number of words (determined in Step 2).

For example:

The causes for urban problems (target case): (excessive factory emissions; massive utilization of fossil fuels; insufficient legislation; construction and demolition dust; vehicle exhaust emissions):

$$\mathrm{Word}\mathrm{vector}\left(\mathrm{target}\mathrm{case}\right)\to v_T=(1,1,0,0,1,1,1,0)$$

The causes of urban problems within the case A: (excessive factory emissions; unfavorable climate conditions; massive utilization of fossil fuels; insufficient legislation);

$$\mathrm{Word}\mathrm{vector}\to v_A=(1,0,1,0,1,1,0,0)$$

Step 4: Calculate similarity using the cosine similarity method.

Based on the previous three steps, the similarity between the target case and the cases in the case library can be calculated using the following formula.

$$Si{m}_{local2}=\cos (\theta )=\frac{V_T\cdot V_A}{\Vert V_T\Vert \Vert V_A\Vert }=\frac{{\displaystyle \sum _{i=1}^n{V}_T{}_i\times V_A{}_i}}{\sqrt{{\displaystyle \sum _{i=1}^n{(V_{Ti})}^2}}\times \sqrt{{\displaystyle \sum _{i=1}^n{(V_{Ai})}^2}}}.$$

(4)

In this context, $V_T\cdot V_A$ represents the dot product of the two vectors, while representing the product of the magnitude (length) of the two vectors. The ratio of $V_T\cdot V_A$ and $Si{m}_{local2}$ represents the similarity of causative factors for urban problems.

3.2.3. Obtaining the Global Similarity

According to the method proposed in this study, the global similarity measurement between the stored case and the target case is determined by a combination of the similarity measurement of urban features and the causes of urban problems. The global similarity can be obtained using the following formula:

$$Si{m}_{global}={\omega }_{local1}\times Si{m}_{local1}+{\omega }_{local2}\times Si{m}_{local2}$$

(5)

where ${\omega }_{local1}$ and ${\omega }_{local2}$ represent the weight of the city features and the causes of urban problems, respectively. Based on expert opinion in the relevant fields, the weight matrix can be presented as $\omega$ = (0.5, 0.5)^T. Then, $Si{m}_{global}$, which denotes the global similarity, can be obtained.

3.3. Case Reuse

In the case reuse process, the case base stores all cases and ranks them based on their global similarity to the target case. In addition, the top five cases are extracted as alternative solutions for users. These alternative solutions from the retrieved cases should be appropriately revised to fit the new situation of the target case. Experts specializing in urban construction and planning will therefore conduct additional research on these alternatives.

3.4. Case Revising

Three strategies, i.e., local rejection, local modification, and local consolidation techniques, are employed to modify the retrieved solutions. Local rejection is the process of eliminating inapplicable solutions for the target city. Local modification stresses modifying a portion of the answer to adapt to the new circumstance. By employing the local consolidation technique, the solutions from several stored cases can be integrated into one. The modified solutions are then bundled into an integrated solution that will be recommended to users.

3.5. Case Retaining

The final step of the SESUPG is case retaining. Through case retaining, the urban problem governance strategies obtained through case reuse and case revising will ultimately become a new part of the case library, combined with inputted urban characteristics and causes of urban problems. The case retain function endows the SESUPG with a self-learning ability that supports the gradual refinement of urban problem governance strategies over time. The case retaining process emphasizes the incorporation of successful experience gained by implementing the solutions to a new problem.

4. Case Study

4.1. Study Area

The case study focuses on verifying the feasibility of the SESUPG, and Wuhan, located in central China’s Hubei Province, was chosen as the target city. Wuhan is a renowned city covering a total area of 8569.15 square kilometers, with a population of 13.65 million [81]. It is located in the subtropical zone on the Jianghan Plain in the middle reaches of the Yangtze River, with a subtropical monsoon climate characterized by abundant rainfall, high temperatures in summer, and mild winters with little rain.

Since 2000, Wuhan has experienced fast urbanization, and by the end of 2021, the urbanization rate had reached 84.56%. Rapid urbanization has significantly boosted economic growth; Wuhan’s gross domestic product (GDP) reached CNY 1771.6 billion in 2021, placing it 9th among the 293 prefecture-level cities in mainland China [81].

However, the rapid urbanization process has also triggered several urban problems, such as traffic congestion, high housing prices, air pollution, etc., which have made Wuhan suffer greatly [82,83,84]. To effectively manage persistent urban problems and maintain the health of Wuhan city, it is imperative to implement appropriate measures. Thus, it is crucial to prioritize the timely diagnosis and tailored prevention of urban problems to ensure sustainable urban health. We took air pollution as an example to demonstrate the application of the SESUPG.

4.2. The Development and Application of SESUPG Software

The Establishment and Application of SESUPG Software

The interface of the Solution Extracted System for Urban Problems Governance (SESUPG) software (V1.0) (Registration No. 2022SR1121882) was developed in MATLAB R2021b, primarily utilizing the App Designer function within the software. The system’s interface facilitates the input of information on urban features, urban problems, and their underlying causes in a user-friendly manner. Moreover, the system provides an interface for displaying retrieved cases that may be of reference for users (see Figure 6).

The internal motivation behind the development of this software follows the sequence outlined in Section 3, as described by the SESUPG. The retrieval results of the software are presented in Figure 7.

4.3. Policy Suggestion and Verification

Policy Suggestion

Upon completion of the retrieval process, the 16 solutions from the top five stored cases were extracted. The extracted solutions underwent a series of strategies, which included local rejection, local modification, and local consolidation, aimed at improving their quality.

(1)

Local rejection

This particular approach was implemented concerning options that were incompatible with the unique circumstances present in Wuhan.

We can take a solution for Sydney as an illustration.

“Strengthen the management patrol of the forest to detect and solve fire problems promptly.”

The topography of Wuhan is predominantly characterized by plains and low hills, with a forest coverage rate of only 14.69% in 2021 [81], as reported in the Wuhan Statistical Yearbook. In contrast, Sydney’s forest coverage rate was 68.70% [4], according to UN data for 2021. Due to this significant difference in forest coverage, the risk of forest fires in Wuhan is relatively lower compared to Sydney. As a result, the proposed solution involving forest fire prevention strategies in Wuhan was deemed unsuitable and, therefore, rejected.

(2)

Local modification

In situations where some of the proposed solutions were not fully compatible with Wuhan’s specific circumstances, local modifications were advised and implemented to tailor the solution to better suit the city’s needs.

“Adopting a regional collaborative management approach. Beijing, Tianjin, and Hebei deepened joint prevention and control of air pollution, established a working mechanism, and set up a working coordination group. Tianjin and the ecological and environmental departments of Beijing and Hebei established a sound collaborative mechanism to promote the continuous improvement of regional ecological and environmental quality.”

Combined with Wuhan’s location characteristics and surrounding city cooperation links, the modified solution was obtained as follows:

“Take a regional collaborative governance approach for controlling air pollutants in Wuhan “1 + 8 Metropolitan Area” such as PM_2.5 and ozone.”

(3)

Local consolidation

Regarding the process of local consolidation, an example has been cited to demonstrate its application. Specifically, this solution has been identified and implemented in three distinct cases.

“Specify the upper limit of air pollutants in respect of critical industry.”

“Propose a low-carbon development strategy, phasing out high-pollution enterprises.”

“The implementation of a comprehensive compliance program for industrial sources of pollution; the full implementation of self-testing and information disclosure for industrial sources of pollution.”

Following a thorough analysis and collaborative design process, the three individual solutions were synthesized and consolidated into a comprehensive integrated solution, outlined as follows.

“Specify the upper limit of air pollutants in critical industry, implement self-testing and information disclosure for industrial sources of pollution, and phase out high-pollution enterprises to propose a low-carbon policy.”

Based on the aforementioned three-step process, a total of four solutions were generated from the top five cases with the most similarity. These solutions were subsequently presented to the Wuhan government for their consideration and reference (see

Table 4. Suitable solutions for Wuhan air pollution.

No	Solution
1	Promote efficient and clean engine technology and the usage of low-pollution energy sources such as natural gas. Propose a low-carbon development strategy
2	Take a regional collaborative governance approach for controlling air pollutants in Wuhan City Ring, such as PM2.5 and ozone.
3	Replace low-capacity, polluting, unsafe, and outdated minibuses with high-capacity, energy-efficient buses. Put in a large number of pure electric buses and cabs, CNG dual-fuel cabs, and LNG buses.
4	Specify the upper limit of air pollutants in critical industries, implement self-testing and information disclosure for industrial sources of pollution, and phase out high-pollution enterprises to propose a low-carbon policy.

).

5. Discussion

In February 2021, the study team presented the proposed ideas to the Wuhan Urban and Rural Construction Bureau. The second and fourth options from Table 4 were adopted and included in the policy report for the “Wuhan Optimized Urban Human Environment Action Plan (2021–2023)”. Furthermore, all four solutions were found to be highly compatible with “The 14th Five-Year Plan for the Ecological Environmental Protection Plan of Wuhan City”, the guiding plan for ecological environment development in Wuhan from 2021 to 2025 [82]. This outcome serves as evidence of the effectiveness of the SESUPG.

The comparison between this study and the previous one has been conducted from two perspectives. Firstly, governments often serve as decision makers for urban problem-solving initiatives, and they frequently employ methods such as the Delphi method and brainstorming to generate solutions. In this process, solutions are often determined by expert opinions or the outcomes of meetings. This approach can lead to decision outcomes that may carry subjectivity, particularly if there is improper expert selection or biases present [83]. Such factors could potentially impact the quality and accuracy of the final recommendations and inevitably introduce subjectivity into decision outcomes. Furthermore, there have been studies through empirical analysis [84,85] focusing on the aim of addressing urban problems by either exploring the evolution patterns of urban problems to mitigate them or by offering targeted recommendations based on comprehensive evaluation outcomes or multi-criteria decision methods (MCDMs) to comprehensively evaluate aspects like urban resilience [24,86] and urban sustainability [87]. However, empirical studies often encounter endogeneity problems with accompanying variables and comprehensive evaluation results tend to be biased, making it challenging to cover a wide range of urban problems. Instead, the SESUPG focuses on seeking governance strategies for urban problems through scientifically quantifiable methods. The SESUPG framework we propose can be likened to a black box, as it does not delve into intricate discussions of the interrelationships among these factors. Instead, it focuses on deriving the most suitable urban problem governance strategies through inputting urban characteristics and the causes of urban problems, thus obtaining appropriate solutions.

Secondly, our research draws a comparison with analogous studies employing the case-based reasoning (CBR) methodology in sustainable management. It is appreciated that prior research has employed the CBR framework in the realm of urban problem governance by reusing past experiences to solve new problems [4,42,43,52,56]. However, existing research has seen limited exploration of the utilization of the CBR framework within the context of urban problem governance. Diverging from past approaches utilizing CBR to address sustainability issues, our approach incorporates the comprehensive consideration of urban features and specific urban problem causes, aimed at generating context-specific urban problem-solving strategies for urban problem governance. This offers more effective governance strategies to urban administrators, thereby enhancing urban sustainability.

6. Conclusions

The phenomenon of rapid urbanization has led to the emergence of various urban problems. However, the complexity of urban problem governance has hindered the development of an effective, location-specific decision-making system to combat these issues. In this study, we proposed an SESUPG method that utilizes successful experience from the past to treat urban problems, and a case study of Wuhan was conducted to verify the feasibility of the approach.

The main contributions of this study are as follows. In terms of the theoretical aspects, we proposed the SESUPG framework based on the CBR method. Compared to previous studies [13,55,60], this framework takes into account both urban features and the causes of urban problems in the process of case representation and case retrieval. This leads to more accurate and comprehensive results in the case retrieval process. In practical terms, we have established a prototype of the SESUPG that can assist government decision makers in quickly retrieving useful experiences as a basis for policy formulation. Furthermore, due to the operating mechanism of this system, it will continuously update and expand its library with ongoing usage. This enables it to provide a wider range of reference and selection options.

However, there are still several limitations that should be appreciated. Firstly, the number of collected cases in this study is limited due to the significant manpower and practical resources required. In further research, more successful cases will be collected using web-crawling techniques. Secondly, the extraction of core etiological terms is complex and difficult to determine. It is prone to errors in the absence of manual review. Therefore, further studies will adopt natural language processing (NLP) or neural network (such as CNN) methods to identify and extract the core etiological terms of urban problem causes.