Promoting Low-Carbonization in the Construction Supply Chain: Key Influencing Factors and Sustainable Practices

Abstract

The construction industry is a major contributor to global carbon emissions, making the decarbonization of its supply chain a critical goal for sustainable development. This study aims to identify key drivers of decarbonization within the construction supply chain and analyze their interrelationships using causal and structural modeling techniques. A bibliometric analysis is conducted to highlight theoretical gaps in the field, followed by the application of the Fuzzy Decision-Making Trial and Evaluation Laboratory (Fuzzy-DEMATEL) method to identify key decarbonization factors, and the Adversarial Interpretive Structural Modeling (AISM) method to construct a causal model that reveals their interactions. The results identify 17 key factors, categorized into five levels, with low-carbon policy guidance and consumer preferences emerging as the most influential. Based on these findings, a three-tier strategy is proposed to guide decarbonization efforts within the construction supply chain. This research provides theoretical insights into decarbonization and offers actionable recommendations for future industry practices and policy development.

1. Introduction

In recent years, the global construction sector has faced an unprecedented decarbonization challenge. According to the United Nations Environment Program (2022), CO₂ emissions from the buildings sector reached 1 billion tons in 2021, accounting for 37% of global energy-related emissions, a 5% increase from the pre-pandemic period. This trend is a significant departure from the goals of the Paris Agreement’s 1.5 °C pathway, which calls for a 50% reduction in building emissions by 2030 [1]. Of particular concern is the significant neglect of the carbon footprint in the materials production chain. Concrete production alone accounts for 8% of global emissions [2], surpassing traditional high-carbon sectors such as shipping (2.9%) and aviation (1.9%) [3]. Furthermore, despite progress in green building design and energy efficiency, construction supply chain activities, including material transportation and waste generation, still account for 11% of global energy-related emissions [4]. With annual emissions growth of 2% since 2015 [5], these figures highlight the fragmentation and limitations of decarbonization improvements in the construction sector.

Structural barriers impede industry transformation. According to a 2020 World Green Building Council (WorldGBC) survey of 1200 construction companies, 68% of decision-makers prioritize material costs over carbon footprint considerations [6]. This cost-oriented thinking permeates supply chain operations, leading developers to focus on short-term returns and contractors to face tight budgetary constraints, with fewer than 40% of developers requesting low-carbon alternatives in tenders [7]. In addition, deeper systemic challenges, such as the lack of harmonized carbon accounting standards at the procurement stage [8], the low adoption rate of life cycle assessment (LCA) methodologies [9], and the inadequacy of policy enforcement mechanisms [10], collectively constitute systemic barriers to decarbonization.

To address the decarbonization challenges in the construction industry, systematically identifying and analyzing key factors and their interrelationships is an effective approach to overcoming existing barriers and facilitating the low-carbon transition of the construction supply chain. Although companies such as Skanska have achieved significant success in real-world projects by applying advanced carbon accounting tools and green building designs, the industry-wide transition continues to face numerous structural challenges. These successful cases demonstrate that scientific carbon management tools and policy incentives can, to some extent, accelerate progress toward low-carbon objectives within the industry [11].

Although this study is innovative in terms of theory and methodology, its scope is primarily limited to specific regions and types of building projects due to data availability and time constraints. As a result, it does not comprehensively cover all regions and types of building projects globally. Additionally, the limited involvement of industry practitioners may affect the in-depth understanding of the practical challenges of decarbonization implementation and constrain the feasibility evaluation of low-carbon strategies in real-world operations. Based on these considerations, this study aims to systematically identify and analyze the key factors influencing the decarbonization of the construction supply chain. Specifically, the study seeks to address two key questions: (1) What are the key factors influencing the decarbonization of the construction supply chain?; (2) What are the relationships between these factors?

Although existing research has made some progress in building supply chain decarbonization, there are still obvious research gaps. Firstly, current research mostly focuses on single factors or macro-level discussions, lacking a systematic analysis of multiple key influencing factors in the process of building supply chain decarbonization. Secondly, the existing literature has limitations in analyzing these influencing factors, especially the over-reliance on the monolithic nature of the DEMATEL-ISM methodology, and the lack of diversified approaches to explore in depth the dynamic relationships between different factors. These limitations prevent existing studies from fully revealing the complexity and multidimensionality of building supply chain decarbonization. Therefore, this study will elaborate on how to fill these research gaps and provide a more comprehensive analytical framework in the methodology section.

2. Review of Related Literature

2.1. Research Trends in Construction Supply Chain Decarbonization

Over recent decades, the environmental impact of the construction industry has been increasingly scrutinized, driving scholarly interest in decarbonizing the construction supply chain. Since 1985, a steady growth in academic publications in this field has been observed, reflecting the sector’s response to global carbon reduction demands, as analyzed using VOSviewer 1.6.20. To ensure that the analysis reflects the most relevant and recent research, studies published between 2005 and 2024 were selected. This timeframe captures recent advancements in decarbonization strategies while excluding older studies with limited relevance to current challenges.

A comprehensive literature review was performed in the Web of Science database with the query ‘T.S. = (Construction supply chain OR Building supply chain OR Low carbon OR Carbon reduction OR Influencing factors)’. An initial set of 18,496 documents was retrieved from this search. Four screening criteria were subsequently applied: (1) relevance to building supply chain decarbonization based on titles and abstracts; (2) restriction to civil engineering and related disciplines; (3) inclusion of English-language studies; and (4) focus on peer-reviewed journal articles. As a result, a refined dataset of 5982 documents was obtained.

As illustrated in Figure 1, the network and heatmap highlight that keywords such as “renewable energy”, “CO₂ emissions”, and “carbon neutrality” dominate the research landscape, reflecting a strong emphasis on energy transition and carbon reduction strategies. However, there is limited focus on “influencing factors”, indicating a gap in comprehensive, integrated analyses specifically addressing the construction supply chain. This gap underscores the need for more in-depth research to explore the unique challenges faced by the construction industry in its pursuit of decarbonization. While the concept of managing construction supply chains began to emerge in the 1990s, this specialized area remains underexplored, particularly regarding its role in supporting low-carbon objectives.

2.2. Review of Research on Low-Carbonization of the Construction Supply Chain

The low-carbonization of the construction supply chain is not only a crucial measure to address global climate change but also has profound significance in promoting the green transformation of the construction industry and reducing carbon emissions. Additionally, it serves as an essential safeguard for achieving sustainable development. Consequently, this issue has garnered widespread attention from the international community in recent years [12]. For example, J. Giesekam [13] explored the challenges of applying alternative materials to reduce carbon emissions within the construction industry by surveying industry professionals and interviewing industry leaders. The study shows that although barriers such as high costs, unclear responsibility allocation, and industry culture exist, these obstacles can be overcome through the early involvement of professionals in the supply chain, effective use of lifecycle costs, and modifications to contracts and tender documents, which could promote the widespread adoption of low-carbon materials. This aligns well with the concept of low-carbon energy and material supply, emphasizing the need to enhance the resilience of low-carbon material supply and promote their broader application.

Furthermore, Y. Yang [14] developed a stochastic game model based on the Moran process to analyze the low-carbon emission strategy as an evolutionarily stable strategy within supplier groups in the construction supply chain. The results highlight that the government can effectively guide suppliers towards low-carbon production through differentiated policies, such as establishing carbon emission standards, offering cost subsidies, implementing reward and punishment measures, and promoting information transparency. This conclusion is closely related to the selection of low-carbon suppliers and contractors, highlighting the critical role of government policies in facilitating the low-carbon transformation.

X. Lai [15] employed system dynamics (SD) methods combined with questionnaires to analyze the driving forces and interrelationships of low-carbon construction technology (LCT) innovation. The study found that the joint involvement of government and private enterprises is crucial to the successful transformation of low-carbon projects. Additionally, the research pointed out that the influence of a single driving factor on the growth of low-carbon projects is limited, whereas system integration plays a pivotal role in achieving low-carbon development.

2.3. Review of Research on Factors Influencing Low-Carbonization of the Construction Industry

Given the current lack of research on factors influencing the low-carbonization of the construction supply chain, this study reviews the relevant research on factors influencing low-carbonization in the construction industry. Q. Du used structural equation modeling (SEM) to propose a hypothetical model that includes “social and government factors”, “market factors”, “technical factors”, and “supply chain coordination factors” to assess the potential for carbon emission reduction in the prefabricated building supply chain (PBSC). The study found that “technical factors” have the greatest impact on reducing carbon emissions in PBSC, while “supply chain coordination factors” have the least impact. Moreover, the research discovered that “market factors” have the most significant indirect influence and highlighted that “low-carbon design levels” and “corporate low-carbon awareness” are key factors in reducing carbon emissions in PBSC [16].

Y. Geng, through a comprehensive literature review and expert interviews, identified six key stakeholders and twenty-one related driving factors in the low-carbon development of the construction industry. The study then applied social network analysis to evaluate the collected survey data. The findings indicated that market demand, competition within the construction industry, and business demands from non-governmental entities are ten key factors significantly influencing the low-carbon building sector. In the participant network, the government, the public, and construction entities occupy central positions and exert significant influence, becoming the key drivers of industry development [17].

Z. Wang, using latent Dirichlet allocation (LDA) topic modeling, identified 21 factors influencing the low-carbon construction behavior (CLCB) of contractors in the Chinese construction industry from relevant literature. The study then used the Decision-Making Trial and Evaluation Laboratory (DEMATEL) to evaluate the causal relationships and centrality of these factors and developed a hierarchical structure using Interpretive Structural Modeling (ISM). The results revealed that “incentive policies for relevant stakeholders” and “low-carbon regulations and supervision” are the main factors influencing low-carbon construction behavior. The study emphasized that low-carbon construction behavior should be guided by policies and subjective awareness, highlighting the role of market and management support in promoting low-carbon construction, with technology as the foundation and economic considerations, directly driving the realization of low-carbon behaviors [18].

2.4. Research Gap Analysis

In conclusion, the above studies provide valuable insights and information for analyzing the factors influencing the low-carbonization of the construction supply chain. However, there are still research gaps that warrant further exploration.

First, according to the results of the bibliometric analysis, the network and heatmaps presented in Figure 1 show that existing studies predominantly focus on keywords such as “renewable energy”, “carbon dioxide emissions”, and “carbon neutrality”. This suggests that these topics have emerged as research hotspots in the field of low-carbon construction. However, due to the limited research directly related to the low-carbonization of the construction supply chain, keywords such as “factors influencing the low-carbonization of the construction supply chain” are scarcely represented in the figure.

Second, the review of research on the low-carbonization of the construction supply chain reveals that while some studies have addressed factor analysis and laid the groundwork for factor extraction in this research, most studies focus on individual factors or macro-level discussions. For instance, some studies concentrate on policy-driven factors, while others emphasize market or technical factors. However, systematic research analyzing the key factors of low-carbonization in the construction supply chain remains relatively scarce.

Finally, the review of research on factors influencing low-carbonization shows that most existing studies employ static analysis methods, such as the DEMATEL-ISM model. Although these methods can reveal the hierarchical relationships between influencing factors, they are unable to dynamically simulate the interactions among factors and their changes over time.

Based on the above analysis, this study adopts an overall supply chain perspective to systematically identify the factors influencing the low-carbonization of the construction supply chain from relevant literature. To determine the importance of each influencing factor and their interrelationships, this study proposes a combined analysis method integrating Fuzzy-DEMATEL and AISM. Compared to traditional DEMATEL and ISM methods, this approach innovates by incorporating fuzzy set theory to address uncertainty, while also employing an antagonistic structure matrix (AISM) to minimize the influence of expert subjective judgment and improve the accuracy of factor hierarchy analysis. Moreover, this method can dynamically optimize the causal relationships among factors, thereby making the research results more robust and scientifically reliable. This approach seeks to address the existing research gap and achieve the study’s objectives.

2.5. Future Research Directions

Building upon the gaps identified in the existing literature, future research could focus on the following areas:

• Quantifying the Impact of Influencing Factors

Future studies should explore both the long-term and short-term impacts of various decarbonization strategies, particularly the effects of policy changes and market fluctuations on construction supply chains. Such research will provide more precise theoretical foundations for the development of decarbonization policies and assist industry practitioners in making informed decisions in response to dynamic market conditions.

2.

Cross-Sector Collaboration for Decarbonization

Decarbonizing construction supply chains requires effective collaboration across multiple sectors. Future research should investigate how industries such as energy, construction materials, and transportation can better coordinate their efforts to accelerate decarbonization. Cross-sector collaboration can maximize decarbonization benefits, thus driving the green transformation of the construction industry.

3.

Tailored Decarbonization Models for Specific Sectors

The construction industry consists of various sectors with distinct decarbonization needs. Future research should focus on developing sector-specific decarbonization models to improve the effectiveness and feasibility of strategies. For example, tailored decarbonization pathways for residential and commercial construction supply chains should be developed, as their energy consumption, material use, and logistics requirements differ. Such models will offer more practical and actionable solutions.

4.

Integration of Advanced Technologies

The integration of advanced technologies, such as artificial intelligence, big data, and machine learning, can significantly enhance the accuracy and predictive capabilities of decarbonization strategies. Future studies should explore how these technologies can be incorporated into decision-making frameworks to improve the analysis and forecasting of decarbonization outcomes, thereby providing scientific support for policy development.

Addressing these research directions will yield more precise and practical solutions for decarbonizing construction supply chains, contributing to a sustainable, low-carbon future for the industry while supporting global emissions reduction goals.

3. Methodology and Factor Analysis

The framework of this study consists of five key stages: conducting a bibliometric analysis and literature review, identifying influencing factors, validating these factors through a questionnaire, performing a Fuzzy-DEMATEL analysis, and conducting an AISM analysis. Each stage includes specific tasks, generating corresponding outputs upon completion (refer to Figure 2 for details).

Stage 1: Literature Review and Theoretical Research Gap Analysis

The literature review identifies gaps in the current research, providing a theoretical foundation for the subsequent stages. The arrow indicates that the literature review supports the identification of research gaps and guides the following stages.

Stage 2: Influencing Factor Extraction

Relevant influencing factors are extracted from the literature, forming an initial list. The arrow connects the identification of research gaps to factor extraction, showing that the identified gaps directly guide the extraction of influencing factors.

Stage 3: Factor Screening and Testing

The identified factors are validated through a questionnaire to ensure their relevance and reliability. The arrow points to factor screening and testing, indicating that the factors are further validated and refined through the questionnaire.

Stage 4: Fuzzy-DEMATEL Analysis

The Fuzzy-DEMATEL method is used to analyze the causal relationships between the factors and generate an impact matrix. The arrow points to Fuzzy-DEMATEL analysis, indicating that the validated factors enter the causal analysis stage to reveal their interrelationships.

Stage 5: AISM Analysis

The results from the previous stages are applied to construct a hierarchical structure model using AISM, providing decision-making support for decarbonization strategies. The arrow connects Fuzzy-DEMATEL analysis to the AISM analysis, indicating that the results from the previous analysis serve as the basis for the AISM model construction.

Each stage’s output serves as the input for the next, ensuring the systematization and progression of the research process.

3.1. Identification of Influencing Factors Through Literature Survey Techniques

To comprehensively analyze the key factors influencing low-carbonization in the construction supply chain, this study employs literature research, field investigations, and questionnaire surveys. Based on the bibliometric analysis, a secondary search was conducted in the Web of Science database using keywords such as “construction supply chain”, “low-carbon development”, “carbon footprint”, and “carbon emission reduction strategies”. Relevant studies on low-carbonization in the construction supply chain were selected for further analysis.

To ensure the scientific rigor and accuracy of the study, the literature selection was based on the following criteria:

• Authority of the Source:

  (1)

  The selected studies were published in high-impact factor academic journals (e.g., Journal of Cleaner Production, Sustainable Cities and Society, Building, and Environment) or by renowned research institutions (e.g., the International Energy Agency (IEA) and the World Green Building Council (WGBC)), ensuring their credibility and reliability;

  (2)

  Journals with high academic impact were prioritized, particularly those classified as JCR Q1 or Q2, and their impact factor (IF) was considered;

  (3)

  All selected literature was indexed in SCI or EI databases to ensure traceability and academic recognition.

• Relevance to Research Content: The title, abstract, and keywords must be highly relevant to low-carbonization in the construction supply chain to ensure specificity.
• Classification of Research Types:

  (1)

  Review Articles: Provide a systematic overview of research progress in low-carbonization within the construction supply chain, forming the theoretical foundation of this study;

  (2)

  Empirical Studies: Utilize real-world data and case analysis to provide quantitative evidence for low-carbon supply chain practices;

  (3)

  Case Studies: Conduct in-depth analysis of specific industries, regions, or enterprises to enhance practical applicability;

  (4)

  Methodological Studies: Develop or optimize models, frameworks, and decision-making tools related to low-carbon supply chain management.

• Academic Influence:

  (1)

  Preference was given to studies with high citation counts or notable academic impact in the fields of low-carbon construction and supply chain management to enhance research representativeness;

  (2)

  The selected literature had to identify and discuss key influencing factors related to low-carbonization in the construction supply chain, such as policies and regulations, green technologies, economic incentives, market demand, and supply chain collaboration, ensuring systematic coverage of the topic.

Based on the criteria above, a total of 18 representative papers were selected, with the following characteristics:

(1)

Average journal impact factor (IF) = 9.6;

(2)

SCI Q1/Q2 journal proportion = 83%;

(3)

Median citation count = 99 (as of 2024).

The selected literature examines five key entities within the construction supply chain—suppliers, developers, designers, contractors, and consumers—while also considering the roles of government, societal context, and inter-firm relationships. Based on this framework, 19 key factors influencing decarbonization in the construction supply chain are identified and classified (as shown in

Table 1. Factors influencing the decarbonization of the construction supply chain.

Perspective	Factor	Sources of Factors	Meaning
Suppliers	Low-carbon energy and materials supply F1	[19,20,21]	Guaranteeing the provision of low-carbon energy, materials, and equipment throughout the construction supply chain to minimize emissions. The transportation of materials and equipment can significantly lower energy consumption and carbon emissions by incorporating low-carbon equipment in downstream supply chain processes.
	Low-carbon equipment supply F2
	Transportation with reduced carbon emissions F3
Builders	Selection of suppliers and contractors with low-carbon practices F4	[22,23,24]	Choosing material and equipment suppliers, general contractors, and subcontractors with strong low-carbon performance by utilizing comprehensive low-carbon screening criteria. Operation and maintenance management focuses on utilizing low-carbon materials and equipment to reduce CO2 emissions.
	Management of properties with low-carbon strategies F5
	Low-carbon construction and operation concept F6
Designers	Low-carbon design assessment tool F7	[25,26,27]	Developing low-carbon design assessment tools helps quantify the environmental benefits of architectural solutions, aligning with the goals of this study.
	Low-carbon structural design F8		Reduce CO2 emissions through structural design of daylighting and lighting, materials, natural ventilation, etc.
	Low-carbon energy design F9		Developing low-carbon energy solutions and implementing sustainable technologies to minimize carbon dioxide emissions.
Constructor	Innovation in low-carbon construction technologies F10	[28,29,30]	Promote innovative low-carbon construction technologies to reduce resource consumption and waste generation. Reduce environmental impacts through construction processes such as abrasion reduction and energy consumption reduction to realize a low-carbon construction process. Reduce carbon emissions during the construction phase through management of systems, technologies, responsibility awareness, employee behavior, and tracking and verification. Recycling of usable construction waste
	Low-carbon construction process F11
	Low-carbon construction management F12
	Low-carbon construction waste recycling F13
Government and society	Low-carbon policy and regulatory guidance F14	[31,32]	Government incentives and carbon emission limitation measures for low-carbon behaviors of supply chain parties
	Low-carbon publicity by social organizations and the media F15		Publicity campaigns by social welfare organizations and the media on the theme of low-carbon.
Consumers	Low-carbon building awareness raising F16	[32,33]	Raising consumer awareness and demand for low-carbon buildings through education and publicity campaigns
	Low-carbon consumption preferences F17		Consumer preference for low-carbon building products or services
Construction supply chain business-to-business	Low-carbon behavioral collaborative management F18	[34,35]	Low-carbon behavior monitoring and collaborative management cooperation among supply chain enterprises
	Low-carbon information sharing F19		Sharing of low-carbon product and technology information among supply chain enterprises

). This systematic review provides a comprehensive foundation for subsequent analysis and ensures high-quality academic support.

3.2. Questionnaire Survey and Statistical Validation

3.2.1. Design and Dissemination of the Survey

The questionnaire was designed to confirm the influencing factors identified in the analysis. Reliability and validity tests were conducted using SPSS 27.0 software to ensure accuracy, while unsuitable factors were removed through T-tests and ANOVA, resulting in a refined system of influencing factors.

To ensure scientific rigor and clarity, the questionnaire was structured to validate the influencing factors identified earlier. The content was divided into three sections:

• Survey Background and Respondent Information: The first section provided an overview of the survey background and objectives while collecting respondents’ basic information, including work experience, educational background, and organizational affiliation;
• Factor Evaluation Using Likert Scale: The second section used a five-point Likert scale, allowing respondents to rate the significance of each factor listed in Table 1, where higher scores reflected a greater impact on the decarbonization of the construction supply chain;
• Feedback and Suggestions: The third section invited respondents to provide feedback on the questionnaire content and suggest improvements.

The questionnaire was distributed to practitioners and researchers in the construction field through both paper and electronic formats. Alumni networks and online platforms, including email and WeChat, were used to ensure broad distribution. A total of 150 questionnaires were distributed (35 paper-based and 115 electronic), with 125 valid responses received (20 paper and 105 electronic), resulting in a response rate of 83.3%.

Table 2. Basic information of respondents of valid questionnaires.

Respondent Demographics	Contexts	Sample Size (Statistics)
Type of organization	Engineering design institute	22
	Construction company	35
	Engineering consulting firm	47
	Scientific research institutes	15
	Other	6
Working experience	Less than five years	35
	Between five to ten years	56
	Between ten to twenty years	25
	More than twenty years	9
Educational background	Below Bachelor’s degree	17
	Bachelor’s degree	53
	Master’s degree	40
	Doctoral degree or Higher	15
Title	Junior title	17
	Intermediate title	54
	Vice-Senior title	33
	Senior title	21

summarizes the basic demographic information of the respondents.

To further validate the identified factors and assess their relative importance, a panel of 15 experts from academia, industry, and policy-making sectors participated in structured interviews and surveys. These expert evaluations provided critical insights into the ranking and classification of influencing factors, ensuring high-quality data for further analysis.

3.2.2. Reliability and Validity Tests

SPSS 27.0 was employed in this study to conduct a reliability analysis, ensuring the validity of the influencing factors presented in Table 1 and evaluating the questionnaire’s consistency. The analysis revealed a high level of internal consistency, with a reliability coefficient of 0.810, confirming the questionnaire’s appropriateness for further investigation.

The values of the Corrected Item–Total Correlation (C.I.T.C.) were analyzed during the reliability test to assess correlations between individual items and the total score. Results revealed that the C.I.T.C. values for ‘Low-carbon Building Awareness Promotion’ and ‘Low-carbon Design Assessment Tool’ were below 0.3, indicating weak correlations with other factors. Consequently, these two factors were excluded to enhance the overall reliability of the analysis. Detailed results are presented in

Table 3. Reliability testing.

Items	Corrected Item–Total Correlation (CITC)	Cronbach’s Alpha If Item Deleted	Cronbach α (Standardized = 0.842)
Low-carbon energy and materials supply	0.387	0.802	0.810
Low-carbon equipment supply	0.444	0.798
Transportation with reduced carbon emissions	0.396	0.801
Selection of suppliers and contractors with low-carbon practices	0.435	0.799
Management of properties with low-carbon strategies	0.432	0.799
Low-carbon construction and operation concept	0.444	0.798
Low-carbon design assessment tool	0.098	0.824
Low-carbon structural design	0.545	0.792
Low-carbon energy design	0.337	0.804
Innovation in low-carbon construction technologies	0.561	0.791
Low-carbon construction process	0.447	0.798
Low-carbon construction management	0.431	0.800
Low-carbon construction waste recycling	0.586	0.793
Low-Carbon Building Awareness Raising	−0.045	0.841
Low-carbon consumption preferences	0.584	0.794
Low-carbon policy and regulatory guidance	0.599	0.793
Low-carbon publicity by social organizations and the media	0.427	0.801
Low-carbon behavioral co-management	0.443	0.800
Low-carbon information sharing	0.442	0.800

.

The validity test results showed a K.M.O. value of 0.823 (exceeding the threshold of 0.5) and a significance level of 0.000 in Bartlett’s test of sphericity, demonstrating statistically significant correlations among the factors. These findings confirm that the data are appropriate for further analysis of factors influencing construction supply chain decarbonization using the Fuzzy-DEMATEL-AISM approach. A detailed summary of the results is presented in

Table 4. Validity test.

KMO		0.823
Bartlett’s Test of Sphericity	Chi-Square	815.280
	$df$	136
	p	0.000

.

3.2.3. Statistical Tests

This study used SPSS software to conduct one-sample t-tests and consistency analysis (using standard deviation as a measure) to assess the importance and consistency of the identified factors. The detailed explanations of each statistical analysis and hypothesis testing are provided below.

To evaluate the importance of each factor, a one-sample t-test was performed. The null hypothesis for the hypothesis test stated that the mean of each factor equals the theoretical midpoint value of 3 (${\mu }_0$ = 3), while the alternative hypothesis suggested that the mean of some factors significantly deviates from 3 [36]. The t-test results indicated that the mean scores for all factors were significantly higher than 3 (with p-values all equal to 0.000), as shown in

Table 5. One-sample t-test results for factor importance assessment.

Items	n	Min	Max	Mean	Std. Deviation	t	p
Low-carbon energy and materials supply	125	1.000	5.000	3.560	0.971	41.010	0.000 **
Low-carbon equipment supply	125	1.000	5.000	3.552	0.979	40.553	0.000 **
Low-carbon transportation	125	2.000	5.000	3.656	0.925	44.167	0.000 **
Low-carbon supplier and contractor selection	125	1.000	5.000	3.472	0.912	42.561	0.000 **
Low-carbon property management	125	1.000	5.000	3.600	0.907	44.378	0.000 **
Low-carbon construction and operation concept	125	1.000	5.000	3.640	0.856	47.558	0.000 **
Low-carbon structural design	125	2.000	5.000	3.704	0.871	47.551	0.000 **
Low-carbon energy design	125	2.000	5.000	3.704	0.833	49.713	0.000 **
Low-carbon construction technology innovation	125	1.000	5.000	3.680	0.938	43.843	0.000 **
Low-carbon construction process	125	2.000	5.000	3.680	0.829	49.636	0.000 **
Low-carbon construction management	125	1.000	5.000	3.528	0.725	54.416	0.000 **
Low-carbon construction waste recycling	125	1.000	5.000	3.488	0.703	55.495	0.000 **
Low-carbon consumption preferences	125	2.000	5.000	3.472	0.630	61.656	0.000 **
Low-carbon policy and regulatory guidance	125	2.000	5.000	3.544	0.678	58.438	0.000 **
Low-carbon publicity by social organizations and the media	125	2.000	5.000	3.712	0.645	64.315	0.000 **
Low-carbon behavioral co-management	125	2.000	5.000	3.688	0.653	63.170	0.000 **
Low-carbon information sharing	125	2.000	5.000	3.704	0.648	63.20	0.000 **

** p < 0.01.

. This suggests that respondents generally considered these factors to be of high importance in the decarbonization process of the construction supply chain. Therefore, these factors were regarded as critical and were retained for further analysis.

To assess the consistency of respondents’ evaluations for each factor, this study used standard deviation as an indicator of data consistency, instead of traditional analysis of variance (ANOVA). Specifically, the null hypothesis for the hypothesis test posited that the variance of the sample scores was close to zero, meaning that the evaluations were highly consistent. The alternative hypothesis suggested that there was significant variance, indicating substantial variation in respondents’ evaluations. For analytical simplicity, standard deviation was chosen as a measure of consistency. If the standard deviation of a factor exceeded 1, it was considered that there was significant variation in evaluations, and such factors were excluded from further analysis.

Based on the analysis results, the standard deviations for all retained factors were found to be less than 1, as shown in Table 5. This indicates that the respondents’ evaluations of the importance of these factors were highly consistent, supporting the alternative hypothesis. Therefore, it can be concluded that respondents’ views on these factors were relatively unified and that there was a high level of consensus.

3.2.4. Factor Coding

Key factors influencing construction supply chain decarbonization were systematically identified based on the above analysis. These factors were categorized according to their relevance to stakeholders’ roles and decision-making processes within the supply chain. To systematically explore their influence on construction supply chain decarbonization, these factors were categorized into seven groups: supply side, builder, designer, constructor, consumer, government, societal level, and inter-company relationships within the supply chain. Each influencing factor was assigned a corresponding code (F1, F2, F3, …, F17), with detailed information in

Table 6. Factor coding.

Name of Influencing Factor	Encodings
Low-carbon energy and material supply	F1
Low-carbon equipment supply	F2
Low-carbon transportation	F3
Low-carbon supplier and contractor selection	F4
Low-carbon property management	F5
Low-carbon construction and operation Concepts	F6
Low-carbon structural design	F7
Low-carbon energy design	F8
Low-carbon construction technology innovation	F9
Low-carbon construction technology	F10
Low-carbon construction management	F11
Low-carbon construction waste recycling	F12
Low-carbon policies and regulations	F13
Low-carbon publicity by social organizations and media	F14
Low-carbon consumption preferences	F15
Low-carbon behavioral collaborative management	F16
Low-carbon information sharing	F17

.

3.3. Fuzzy-DEMATEL-AISM-Based Factor Analysis

3.3.1. Introduction to Integrated Methodology

To effectively analyze the key factors and their interrelationships in the decarbonization of the construction supply chain, this study adopts an integrated analytical framework that combines Fuzzy-DEMATEL with Adversarial Interpretive Structural Modeling (AISM). This integrated method aims to overcome the limitations of traditional methods when addressing complex issues in the construction industry’s decarbonization. By combining fuzzy mathematics and adversarial learning, it not only reveals causal relationships between factors but also enhances the robustness and adaptability of the model in uncertain environments. The following sections elaborate on the applicability of these two methods in the field of construction decarbonization and explain the advantages of their combination:

• The Suitability of Fuzzy-DEMATEL for Building Decarbonization

The decarbonization of the construction supply chain involves the interaction of various factors across multiple dimensions, including policies, technologies, market demands, and social behaviors, all of which exhibit high uncertainty. In this context, Fuzzy-DEMATEL is particularly appropriate. Fuzzy-DEMATEL integrates fuzzy set theory [37] with the Decision-Making Trial and Evaluation Laboratory (DEMATEL) method [38] to effectively address the uncertainty and fuzziness of information. In the context of building decarbonization, Fuzzy-DEMATEL helps extract qualitative insights from expert judgments and converts these qualitative assessments into quantitative data through fuzzy processing, thus uncovering the causal relationships between key factors [39].

For example, the relationship between low-carbon technologies and policies in the construction industry is fraught with uncertainty. Fuzzy-DEMATEL enables experts to assess uncertainties based on experience while minimizing human error [40]. This feature makes Fuzzy-DEMATEL especially suitable for the field of building decarbonization, as it can reveal the complex interactions between policies, technologies, and market demands, accurately identifying the key drivers of decarbonization.

2.

The Applicability of AISM for Building Decarbonization Research

In contrast to Fuzzy-DEMATEL’s qualitative analysis strengths, AISM (Adversarial Interpretive Structural Modeling) enhances the dynamism and robustness of the analysis by introducing an adversarial learning mechanism. While traditional Interpretive Structural Modeling (ISM) methods can identify causal relationships, they have limited capability in handling hierarchical relationships and dynamic changes within complex systems [38]. AISM, by simulating adversarial learning mechanisms [41], allows the model to dynamically adjust based on data changes, thereby identifying more precise and refined causal relationships.

Specifically, for building decarbonization, AISM can adapt to the changing policy environment, technological advancements, and market demands. Through adversarial learning, it continuously optimizes model outcomes. For example, AISM can identify which policy measures are most influential in the decarbonization process and which technological innovations are the strongest drivers, providing more comprehensive and detailed decision support [42]. This dynamic adjustment capability makes AISM an ideal choice for analyzing key factors and their hierarchical relationships in the building decarbonization process.

3.

The Advantages of Combining Fuzzy-DEMATEL and AISM

Although Fuzzy-DEMATEL and AISM each offer distinct advantages, their combination provides a stronger analytical capability. Fuzzy-DEMATEL offers a clear causal relationship framework, converting expert qualitative judgments into quantitative data while effectively managing uncertainty. AISM, on the other hand, further refines causal relationships and hierarchical structures through adversarial learning, enhancing the model’s adaptability to complex and dynamic environments.

By combining Fuzzy-DEMATEL and AISM, we can not only uncover the causal relationships among key factors in construction supply chain decarbonization but also dynamically adjust the model based on data changes, ensuring the efficiency and accuracy of the analysis [43]. For example, in the decision-making process of building decarbonization, Fuzzy-DEMATEL provides an initial causal relationship framework, while AISM continuously optimizes this framework, making the interrelations between factors more precise and, therefore, providing more reliable decision support for policymakers and enterprises.

Furthermore, the combination of Fuzzy-DEMATEL and AISM offers additional advantages:

(1)

Increased Transparency of Results: Fuzzy-DEMATEL clarifies causal relationships through fuzzy processing, while AISM further optimizes and adjusts, strengthening the robustness of the analysis results;

(2)

Enhanced System Adaptability: The adversarial learning mechanism in AISM allows the model to optimize based on real-time data, which is crucial when dealing with dynamic factors such as policy changes and technological innovations in building decarbonization;

(3)

Integration of Qualitative and Quantitative Analyses: Fuzzy-DEMATEL combines expert qualitative judgment with quantitative data, while AISM optimizes this combination through adversarial learning, ensuring high alignment between theoretical and practical applications of the analysis.

The Fuzzy-DEMATEL-AISM method provides a robust framework for analyzing factor hierarchies and causal relationships, offering critical insights into construction supply chain decarbonization. The methodological workflow is illustrated in Figure 3, which highlights the key stages of the Fuzzy-DEMATEL and AISM analysis.

Stage 1: Initialization of the Impact Matrix

The process begins with the initialization of the impact matrix, where the relationships between various influencing factors are outlined. The arrow points to “Fuzzy Number Conversion”, indicating that the initial impact matrix is transformed into fuzzy numbers to facilitate further analysis.

Stage 2: Fuzzy Number Conversion

The initialized impact matrix is converted into fuzzy numbers, representing uncertainty and providing flexibility in the analysis. The arrow connects “Fuzzy Number Conversion” to “Defuzzification”, signifying the transition from fuzzy numbers to more precise data through defuzzification.

Stage 3: Defuzzification

In this stage, fuzzy numbers are defuzzified to produce crisp, precise values, which are then used for further analysis. The arrow points to “Normalization”, indicating that the defuzzified values undergo normalization to ensure consistency and comparability across factors.

Stage 4: Normalization

Normalization standardizes values to a common scale, ensuring that all factors can be effectively compared. The arrow connects “Normalization” to “Comprehensive Impact Matrix”, reflecting the final step where the normalized values are consolidated into a comprehensive matrix for causal analysis.

Stage 5: AISM Analysis

The comprehensive impact matrix is then used in AISM analysis to construct a hierarchical structure that illustrates the relationships between the factors. The arrow from “Comprehensive Impact Matrix” to “AISM Analysis” shows how the final matrix serves as the foundation for AISM model construction, providing valuable decision-making support for decarbonization strategies.

Each stage of the Fuzzy-DEMATEL-AISM method is interconnected, with the outputs of one stage providing the necessary input for the next, ensuring a systematic and logical progression of the analysis.

3.3.2. Construction of Influencing Factor Indicators

The factors impacting the decarbonization of the construction supply chain were identified using literature reviews, expert consultations, and field research. These influencing factors are represented as follows:

$$G=\left\{g_1,g_2,g_3\dots g_n\right\}$$

(1)

where $g_i$ represents each individual factor. These factors were selected based on their relevance and significance in achieving low-carbon objectives.

3.3.3. Expert Assessment of Influencing Factors

To ensure the accuracy and scientific validity of this study, a panel of 15 experts from the construction industry was invited to assess the identified influencing factors. The experts were selected based on the following key criteria:

• Professional Background and Extensive Experience: The selected experts come from academia, industry, and policy-making sectors, with extensive experience and in-depth knowledge of decarbonization within the construction industry. This diversity ensures that the decarbonization influencing factors in the construction supply chain are assessed from multiple perspectives, thereby ensuring a comprehensive and accurate evaluation;
• Industry Representation: The experts’ backgrounds span construction design, construction, supervision, engineering consulting, and academic research. This diverse range of expertise allows the expert panel to cover all aspects of the construction supply chain. The experts’ practical work experience provides an objective and comprehensive evaluation of the key factors and their interrelationships in the decarbonization process;
• Rich Work Experience: The selected experts have work experience ranging from 5 to 15 years, as well as over 15 years, covering different career stages. This ensures both the comprehensiveness and depth of the evaluation. It also prevents bias by including experts with varying levels of experience, from highly experienced to moderately experienced;
• Academic Background and Technical Expertise: Most of the experts hold doctoral degrees or higher, possessing strong theoretical and technical capabilities. This high level of academic training enables them to accurately understand the complex influencing factors of decarbonization and provide valuable insights during the assessment process;
• Professional Title and Authority: The expert panel includes senior and associate senior professionals with significant authority in their respective fields. Their professional stature contributes to the authority and reliability of the evaluations provided for the study.

The detailed background information of the experts can be found in

Table 7. Background information of experts.

	Contexts	Sample Size (Statistics)
Nature of the Organization	Engineering Design Institute	3
	Constructor	2
	Supervisory Unit	2
	Engineering Consulting Unit	2
	Universities and Research Units	6
Working Experience	5–15 Years	3
	More Than 15 Years	12
Academic Background	Graduate Degree	4
	Doctoral Degree or Higher	11
Title	Associate Senior	4
	Senior	11

. Based on this background, the experts conducted a professional assessment of the interrelationships between the influencing factors, according to the pre-established evaluation criteria (

Table 8. Semantic conversion table.

Expert Evaluation	Impact Value	Trigonometric Fuzzy Number
No Influence	0	(0, 0, 0.1)
Negligible Effect	1	(0.1, 0.3, 0.5)
Slight Impact	2	(0.3, 0.5, 0.7)
Moderate Effect	3	(0.5, 0.7, 0.9)
Significant Influence	4	(0.7, 0.9, 1.0)
Profound Impact	5	(0.9, 1.0, 1.0)

). This expert evaluation ensures a comprehensive and scientifically rigorous analysis of the decarbonization factors in the construction supply chain.

The evaluation matrix was constructed as$B={\left|b_{ij}\right|}_{n^{*}n}$, where $b_{ij}$represents the degree of direct influence of factor$g_i$on factor $g_j$.

The criteria for scoring the impacts were as follows:

• No Influence: “0”;
• Negligible Effect: “1”;
• Slight Impact: “2”;
• Moderate Effect: “3”;
• Significant Influence: “4”;
• Profound Impact: “5”.

Expert opinions were obtained through face-to-face and telephone interviews to enhance the reliability and scientific rigor of the assessments. The collected data were processed using MATLAB R2023a and Python 3.12.0, ensuring consistency and accuracy. Key metrics, such as influence scores and centrality, were calculated to determine factor hierarchies, which served as the foundation for further analysis.

This thorough expert evaluation approach ensured that the collected data were of high quality, providing a robust basis for subsequent DEMATEL computations in Section 3.3.6.

3.3.4. Fuzzification of Expert Evaluations

The direct influence matrix obtained from expert assessments was transformed into a triangular fuzzy number matrix using a semantic conversion table (see Table 7). A triangular fuzzy number $\tilde{N}$ is represented by the triplet $(l,m,r)$, where

• l: lower limit;
• m: most probable value;
• r: upper limit.

The membership function ${\mu }_{\tilde{N}}\left(x\right)$ is expressed as follows:

$${\mu }_{\tilde{N}}(x)=\left\{\begin{array}{cc}(x-l)/(m-l)& l⩽x⩽m\\ (r-x)/(r-m)& m⩽x⩽r\\ 0& x>r \mathrm{or} x<l\end{array}\right.$$

(2)

The fuzziness of $\tilde{N}$ is determined by $r-l$, with larger values indicating higher uncertainty. Through this fuzzification process, ten fuzzy direct influence matrices were generated, addressing uncertainties in the expert evaluations and providing a comprehensive representation of the inter-factor relationships.

3.3.5. Defuzzification Using the C.F.C.S. Method

In this study, the C.F.C.S. (Converting Fuzzy Data into Crisp Scores) method [44] was used to convert the fuzzy direct influence matrix into a clear, crisp form. This approach enhances the accuracy of factor analysis by reducing the influence of subjectivity while maintaining the integrity of the numerical data. The C.F.C.S. method calculates the upper and lower bounds for the scores using fuzzy minima and maxima. Subsequently, the total score is obtained by computing the weighted average of the membership functions, minimizing any loss of precision.

The specific steps are as follows:

• Standardization:

$$x{l}_{ij}^k={\displaystyle \frac{l_{ij}^k-min{l}_{ij}^k}{{\mathsf{\Delta }}_{min}^{max}}},$$

(3)

$$x{m}_{ij}^k={\displaystyle \frac{m_{ij}^k-min{l}_{ij}^k}{{\mathsf{\Delta }}_{min}^{max}}},$$

(4)

$$x{r}_{ij}^k={\displaystyle \frac{r_{ij}^k-min{l}_{ij}^k}{{\mathsf{\Delta }}_{min}^{max}}},$$

(5)

${\mathsf{\Delta }}_{min}^{max}=max{r}_{ij}^k-min{l}_{ij}^k$, Standardization: The lower, middle, and upper bounds of the fuzzy numbers are normalized to facilitate comparability across different factors.

2.

Calculation of standardized values for the left and right sides:

$$xl{s}_{ij}^k={\displaystyle \frac{x{m}_{ij}^k}{1+x{m}_{ij}^k-x{l}_{ij}^k}},$$

(6)

$$xr{s}_{ij}^k=x{r}_{ij}^k/(1+x{r}_{ij}^k-x{m}_{ij}^k),$$

(7)

3.

Calculation of the total standardized value:

$$x_{ij}^k={\displaystyle \frac{\left[xl{s}_{ij}^k\left(1-xl{s}_{ij}^k\right)+xr{s}_{ij}^{k}xr{s}_{ij}^k\right]}{\left[1-xl{s}_{ij}^k+xr{s}_{ij}^k\right]}},$$

(8)

4.

The defuzzification value for each expert’s evaluation denoted as $k$ is then computed:

$$z_{ij}^k=min{l}_{ij}^k+x_{ij}^k{\mathsf{\Delta }}_{min}^{max},$$

(9)

5.

Combining the evaluations of x experts yields the defuzzified direct impact matrix:

$$z_{ij}={\displaystyle \frac{1}{p}}\left(z_{ij}^1+z_{ij}^2+\cdots +z_{ij}^p\right),$$

(10)

Defuzzification: Calculate a precise score for each factor by integrating expert evaluations into a unified matrix.

Several key factors influencing the decarbonization of the construction supply chain were identified, including F1, F2, F3, …, and F17 (see Table 3). A scale of 0 to 5, ranging from “No impact” to “very high impact”, was used to quantify the interactions among these factors. Data on factor interactions, particularly the strength of their mutual influence, were collected through expert surveys.

This methodology ensured that the resulting matrix accurately captured the interrelationships within the construction supply chain. The data derived reflected the degree of influence and relevance of each factor in the actual construction supply chain context. While the study emphasized the relative strength of factor interactions, it did not distinguish the directional effects of these influences. To mitigate the potential bias, the arithmetic mean of expert evaluations was calculated, resulting in a precise initial direct influence matrix (as shown in Figure 4).

3.3.6. DEMATEL-AISM Analysis

To further analyze the causal relationships among factors, the DEMATEL method was employed. The structured expert evaluation results (detailed in Section 3.3.3) were converted into a direct-relation matrix, and MATLAB/Python tools were used to conduct DEMATEL calculations. These tools ensured high computational accuracy and allowed for the precise identification of key influencing factors.

• Normalized influence matrix.

Various methods exist for matrix normalization, and in this study, the row maximum method [45] is applied. The directed topology map construction follows. This approach involves identifying the largest value within each row of the matrix$A$, Once the maximum value for a row is determined, all elements of the matrix $Z$ in that row are divided by this value. As a result, the standardized influence matrix $B$ is constructed, ensuring the comparability of the matrix’s elements across rows.

$$B={\displaystyle \frac{x_{ij}}{\max \left({\sum }_{j=1}^n{x}_{ij}\right)}},$$

(11)

The calculations are shown in Figure 5.

2.

Integrated impact matrix.

The integrated system matrix captures the combined effects of the interactions among the different elements within the system.

$$T=\left(B+B^2+\cdots +B^k\right)={\sum }_{k=1}^{\mathrm{\infty }}{B}^k=B{\left(I-B\right)}^{-1},$$

(12)

where $I$ is the unit matrix. This is shown in Figure 6.

3.

The measures of influence degree, centrality, causality, and weight for each factor are systematically calculated.

(1)

Degree of Influence: Influence Degree: Defined as the total impact of an individual element on all other elements; this measure is computed by summing the values of each row in the matrix $T$. Denoted as $D_i$, it quantifies the extent to which a specific element influences the system.

$$D_i={\sum }_{j=1}^n{x}_{ij},\left(i=\mathrm{1,2},\dots ,n\right),$$

(13)

(2)

Degree of Influence: This term refers to the sum of all columns in the matrix $T$, which indicates the total influence that the elements in each column exert on all other elements within the system. The value is represented as $C_i$.

$$C_i={\sum }_{j=1}^n{x}_{ji},\left(i=\mathrm{1,2},\dots ,n\right),$$

(14)

(3)

Centrality: Centrality reflects the importance and positional relevance of a factor within the evaluation system. This metric, denoted as $M_i$, is derived by aggregating the total influence exerted by an element and the combined influence it receives from all other elements. It provides a comprehensive measure of the role and significance of the element in the overall system.

$$M_i=D_i+C_i,$$

(15)

(4)

Degree of Causality: The causality degree of an element is obtained by calculating the difference between the total influence received from other elements and the influence it exerts on them. Represented as $R_i$, this value reflects the element’s net causative role in the evaluation framework.

$$R_i=D_i-C_i,$$

(16)

Normalization and Weighting: The centrality values are then normalized, which allows the calculation of indicator weights. The outcomes of this process are presented in

Table 9. Degrees of influence, influenced, centrality, and cause.

Factor	Influence	Rank	Influenced	Rank	Causality	Rank	Center	Rank
F1	0.864	14	1.695	1	−0.831	17	2.559	6
F2	0.78	15	1.504	7	−0.724	14	2.284	17
F3	0.767	16	1.535	6	−0.768	15	2.302	16
F4	1.303	8	1.613	2	−0.31	11	2.916	2
F5	0.754	17	1.559	3	−0.805	16	2.313	15
F6	1.415	5	1.243	11	0.172	5	2.658	3
F7	1.062	12	1.415	8	−0.353	12	2.477	12
F8	1.175	10	1.343	10	−0.168	9	2.518	8
F9	1.701	4	1.541	4	0.16	6	3.242	1
F10	1.115	11	1.398	9	−0.283	10	2.513	9
F11	1.294	9	1.241	12	0.053	8	2.535	7
F12	0.916	13	1.54	5	−0.624	13	2.456	13
F13	2.149	1	0.168	17	1.981	1	2.317	14
F14	1.776	3	0.848	15	0.928	3	2.624	4
F15	1.823	2	0.664	16	1.159	2	2.487	11
F16	1.33	7	1.174	13	0.156	7	2.504	10
F17	1.413	6	1.157	14	0.256	4	2.57	5

.

In the Cartesian coordinate system, centrality is plotted along the horizontal axis, and causality is represented on the vertical axis. This graphical representation (see Figure 7) provides a clearer understanding of each factor’s role in the low-carbon building supply chain and offers guidance for practical implementation.

Factors with high driving force and low dependence, such as F13 (low-carbon policy and regulatory guidance), F15 (low-carbon consumer preferences), and F16 (low-carbon advocacy by social organizations and media), are located in the upper-right quadrant. These factors are key drivers in the framework and play a crucial role in promoting the overall low-carbon transformation of the industry. For instance, F13 (low-carbon policy and regulatory guidance) provides the necessary regulatory framework and incentives to guide the market and industry towards low-carbon objectives. F15 (low-carbon consumer preferences) and F16 (low-carbon advocacy) help create market demand and societal support for low-carbon buildings by raising public awareness. Therefore, prioritizing interventions for these factors is crucial. Strengthening policy support, increasing consumer demand for low-carbon products, and expanding public participation through social advocacy will accelerate the decarbonization process and create a stable foundation for the transition.

On the other hand, factors with low driving force and high dependence, such as F2 (low-carbon equipment supply), F3 (low-carbon transportation), F7 (low-carbon structural design), F10 (low-carbon construction process), and F12 (low-carbon construction waste recycling), are located in the lower-left quadrant. These factors are highly influenced by external drivers, particularly government policies and market demand. In practice, management efforts should focus on stabilizing and strengthening the key drivers that influence these factors, such as F13 (policy) and F15 (consumer demand), to ensure the optimal outcomes of these dependent factors. For example, F7 (low-carbon structural design) and F3 (low-carbon transportation) should be optimized by reinforcing low-carbon technology policies and promoting green procurement in the market.

Factors in the autonomous quadrant, such as F1 (low-carbon energy and material supply), F4 (low-carbon supplier and contractor selection), F5 (low-carbon property management), F6 (low-carbon construction and operation concepts), and F11 (low-carbon information sharing), exhibit low driving force and low dependence, meaning that their impact on the overall system is limited. Although these factors are more independent, they may pose potential risks and instability in certain situations. Therefore, even though their direct impact on decarbonization is smaller, continuous monitoring is important to prevent potential disruptions caused by their independence.

In conclusion, the graphical analysis reveals the practical applications and intervention strategies for each factor. F13 and F15, identified as major drivers, should be prioritized for targeted intervention and policy support, while factors in the autonomous quadrant, such as F1 and F5, require regular monitoring to prevent potential risks from their independence and maintain overall system stability.

4.

Establishment of adjacency matrix.

To establish the adjacency matrix $A$, the internal relationships between influencing factors were determined through expert scoring. An evaluation method based on binary relations was applied, where $f_{ij}$ = 1 indicates a direct influence between factors, and $f_{ij}$ = 0 indicates no such influence, as shown in Equation (17).

$$fij\left\{\begin{array}{c}1 There is a binary relation for{ f}_{ij}\\ 0 No binary relation to f_{ij}\end{array}\right.,$$

(17)

The threshold value φ, φ $\in \left[\mathrm{0,1}\right]$, is a critical parameter that determines the inclusion of relationships in the adjacency matrix. When sample data are limited, φ can be set to 0 based on expert judgment. In this study, the values were refined through multiple iterations, utilizing the mean $\left(\alpha \right)$ and standard deviation $\left(\beta \right)$ of the integrated impact matrix to ensure the analysis adhered to rigorous scientific standards. Following this approach, the adjacency matrix $A$ was established to represent the direct influence relationships between factors, as shown in

Table 10. Adjacency matrix A.

	F1	F2	F3	F4	F5	F6	F7	F8	F9	F10	F11	F12	F13	F14	F15	F16	F17
F1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F4	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F5	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F6	1	0	0	0	1	0	0	0	1	0	0	1	0	0	0	0	0
F7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F9	0	0	0	1	0	1	0	0	0	1	1	1	0	1	0	0	0
F10	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F11	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0
F12	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F13	1	1	1	1	1	1	1	1	0	1	1	1	0	0	0	1	1
F14	1	1	1	1	1	1	1	1	0	1	1	1	0	0	0	0	0
F15	1	1	1	1	1	1	1	1	1	0	0	1	0	0	0	1	1
F16	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F17	1	1	1	0	1	0	0	0	1	0	0	0	0	0	0	0	0

.

5.

Calculate the reachable matrix

The reachable matrix $M$ was derived by iteratively multiplying the adjacency matrix $A$ with the unit matrix $I$. The calculation continues until the matrix stabilizes, indicating no further changes, as described in Equation (18). The finalized reachable matrix $M$ (Table 10) reveals all possible paths between factors, indicating the degree of interconnection and dependency within the system.

The calculation of the reachable matrix $M$ is formalized as follows:

$${\left(\mathrm{A}+\mathrm{I}\right)}^{\mathrm{K}-1}\ne {\left(\mathrm{A}+\mathrm{I}\right)}^{\mathrm{K}}={\left(\mathrm{A}+\mathrm{I}\right)}^{\mathrm{K}+1}=\mathrm{M},$$

(18)

The reachable matrix $M$ is obtained from Equation (2) (

Table 11. Reachable matrix M.

	F1	F2	F3	F4	F5	F6	F7	F8	F9	F10	F11	F12	F13	F14	F15	F16	F17
F1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F2	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F3	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F4	1	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0
F5	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0
F6	1	1	1	1	1	1	1	1	1	1	1	1	0	1	0	0	0
F7	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0
F8	1	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0
F9	1	1	1	1	1	1	1	1	1	1	1	1	0	1	0	0	0
F10	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0
F11	0	0	0	0	0	0	0	0	0	0	1	1	0	0	0	0	0
F12	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0
F13	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	1	1
F14	1	1	1	1	1	1	1	1	1	1	1	1	0	1	0	0	0
F15	1	1	1	1	1	1	1	1	1	1	1	1	0	1	1	1	1
F16	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0
F17	1	1	1	1	1	1	1	1	1	1	1	1	0	1	0	0	1

).

6.

Creating a general skeleton matrix

The reachable matrix$M$ undergoes a “shrinking” process, where points forming loops are compressed into a single point. After this operation, the reachable matrix is updated to $M^{\prime }$. Subsequently, edge reduction is performed, which effectively eliminates redundant paths. The method for this operation is as follows:

$${\mathrm{S}}^{\prime }={\mathrm{M}}^{\prime }-{\left({\mathrm{M}}^{\prime }-\mathrm{I}\right)}^2-\mathrm{I},$$

(19)

The matrix $M^{\prime }$ is subjected to edge reduction, resulting in a skeleton matrix ${\mathrm{S}}^{\prime }$. By substituting loop elements, the generalized skeleton matrix $\mathrm{S}$ is obtained, as shown in

Table 12. Generalized skeleton matrix S.

	F1	F2	F3	F4	F5	F6	F7	F8	F9	F10	F11	F12	F13	F14	F15	F16	F17
F1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F4	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F5	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F6	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F9	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F10	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F11	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0
F12	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F13	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1
F14	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F15	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1
F16	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
F17	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

.

7.

Hierarchical Extraction

Using the reachability matrix, we calculate the prior set $\mathrm{Q}\left(e_i\right)$, reachable set $\mathrm{R}\left(e_i\right)$, and the intersection set $\mathrm{J}\left(e_i\right)$ (as shown in

Table 13. Analysis of reachable sets, predecessor sets, and their overlapping elements.

Elemental Indicators	$\mathbf{Reachable} \mathbf{Collections} \mathbf{R}$	$\mathbf{Pre}-\mathbf{Package} \mathbf{Q}$	Intersection (Symbol ∩) (Set Theory) J
F1	[1]	[1, 4, 6, 8, 9, 13, 14, 15, 16, 17]	[1]
F2	[2]	[2, 6, 9, 13, 14, 15, 16, 17]	[2]
F3	[3]	[3, 6, 9, 13, 14, 15, 16, 17]	[3]
F4	[1, 4]	[4, 6, 9, 13, 14, 15, 17]	[4]
F5	[5]	[5, 6, 9, 13, 14, 15, 17]	[5]
F6	[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14]	[6, 9, 13, 14, 15, 17]	[6, 9, 14]
F7	[7]	[6, 7, 9, 13, 14, 15, 17]	[7]
F8	[1, 8]	[6, 8, 9, 13, 14, 15, 17]	[8]
F9	[1, 2,3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14]	[6, 9, 13, 14, 15, 17]	[6, 9, 14]
F10	[10]	[6, 9,10, 13, 14,15,17]	[10]
F11	[11, 12]	[6, 9, 11, 13, 14, 15, 17]	[11]
F12	[12]	[6, 9, 11, 12, 13, 14, 15, 17]	[12]
F13	[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 16, 17]	[13]	[13]
F14	[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14]	[6, 9, 13, 14, 15, 17]	[6, 9, 14]
F15	[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 17]	[15]	[15]
F16	[1, 2, 3, 16]	[13, 15, 16]	[16]
F17	[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 17]	[13, 15, 17]	[17]

). While the reachable and prior sets illustrate the relationships between elements, they do not directly reveal the hierarchical structure. Therefore, we need to divide the hierarchy by the relationship between these three. If the intersection set is equal to the reachable set, the element is the first layer factor (UP type); if the intersection set is equal to the prior set, it is the penultimate layer factor (DOWN type). After that, the stratified elements are deleted and the process is repeated until the stratification is completed.

$$\mathrm{J}\left({\mathrm{e}}_{\mathrm{i}}\right)=\mathrm{R}\left({\mathrm{e}}_{\mathrm{i}}\right)\hspace{1em}\mathrm{U}\mathrm{P} \mathrm{T}\mathrm{o}\mathrm{p}\mathrm{o}\mathrm{l}\mathrm{o}\mathrm{g}\mathrm{y} \mathrm{H}\mathrm{i}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{r}\mathrm{c}\mathrm{h}\mathrm{y} \mathrm{E}\mathrm{x}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n},$$

(20)

$$\mathrm{J}\left({\mathrm{e}}_{\mathrm{i}}\right)=\mathrm{Q}\left({\mathrm{e}}_{\mathrm{i}}\right)\hspace{1em}\mathrm{D}\mathrm{O}\mathrm{W}\mathrm{N} \mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}\mathrm{l}\mathrm{o}\mathrm{g}\mathrm{y} \mathrm{l}\mathrm{e}\mathrm{v}\mathrm{e}\mathrm{l} \mathrm{e}\mathrm{x}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n},$$

(21)

Based on the data presented in the previous table, adversarial cascade extraction is conducted, and the resulting findings are shown in

Table 14. Confrontation hierarchy.

UP Topology-Based (Result-First) Hierarchical Extraction Results		DOWN Topology-Based (Cause-First) Hierarchical Extraction Results
Layer	Key Constituent	Layer	Key Constituent
Tier 1 factors	[1, 2, 3, 5, 7, 10, 12]	Penultimate Tier 5 Factor	[1, 12]
Tier 2 factors	[4, 8, 11, 16]	Penultimate Tier 4 Factors	[2, 3, 4, 5, 7, 8, 10, 11]
Tier 3 factors	[6, 9, 14]	Penultimate 3 factors	[6, 9, 14]
Tier 4 factors	[17]	Penultimate factor	[16, 17]
Tier 5 factors	[13, 15]	Penultimate 1 factor	[13, 15]

.

8.

Mapping of Directed Topology Hierarchies

A directed topology map was constructed based on hierarchical relationships derived from the skeleton matrix and the hierarchical table. The UP-type hierarchy represents a bottom-up flow in which lower-level factors provide foundational support for upper-level factors. In contrast, the DOWN-type hierarchy represents a top-down flow, with higher-level factors exerting direct influence and providing feedback to foundational elements.

In this diagram:

• Directed line segments represent the reachable relationships among factors;
• Bi-directional arrows indicate loops that highlight mutual influence or feedback;
• Lower-level factors indicate fundamental contributions, whereas upper-level factors exert direct effects. Intermediate factors exhibit mutual interdependence and serve transitional roles within the hierarchy.

Figure 8 illustrates the UP-type and DOWN-type directed topology diagrams, comprehensively visualizing hierarchical relationships and factor interdependencies. This graphical representation clarifies the structure and interactions of influencing factors, facilitating an understanding of complex systems such as construction supply chain decarbonization.

3.4. Sensitivity Analysis

To test how changes in expert weight distribution affect the results, this study compares multiple scenarios to assess the stability of the system (see

Table 15. Comparison of sensitivity analysis results under different expert weighting scenarios.

Scenario 1 (Increased Weight for Senior Experts)				Scenario 2 (Increased Weight for Academic Experts)
Factor	Influence Fluctuation	Cause Fluctuation	Hierarchical Stability	Factor	Impact Fluctuation	Cause Fluctuation	Hierarchical Stability
F13	+0.3% (2.155)	+0.6% (+1.993)	Cause Layer (Unchanged)	F13	−0.3% (2.142)	−0.7% (+1.968)	Cause Layer (Unchanged)
F15	−0.4% (1.816)	0.0% (+1.159)	Cause Layer (Unchanged)	F15	+0.4% (1.830)	0.0% (+1.159)	Cause Layer (Unchanged)
F9	+0.6% (1.712)	+12.5% (+0.18)	Cause Layer (Unchanged)	F9	−0.4% (1.694)	−6.3% (+0.15)	Cause Layer (Unchanged)
F4	−0.4% (1.298)	+3.2% (−0.32)	Intermediate Layer (Unchanged)	F4	+0.4% (1.308)	−6.5% (−0.29)	Intermediate Layer (Unchanged)
F12	−0.8% (0.909)	+2.4% (−0.639)	Outcome Layer (Unchanged)	F12	+0.8% (0.923)	−2.4% (−0.609)	Outcome Layer (Unchanged)
Global Statistics		Average fluctuation: ±0.5%		Average fluctuation: ±4.1%		85% factor hierarchy stability

). The baseline scenario assumes equal weights for all experts. Two additional scenarios were tested: Scenario 1 increased the weight of senior experts (with more than 15 years of experience) by 50%, while Scenario 2 increased the weight of academic experts (with PhDs or higher) by 20%. This allows us to examine how the perspectives of industry-experienced and academic experts impact the model.

Regarding factor selection, three main factors were identified based on their role in the system (see Table 15). The root cause layer includes the most influential variables, such as F13 (low-carbon policies) and F15 (low-carbon consumption preferences), which have a fundamental impact on the system. The intermediary layer includes key technical factors, such as F9 (low-carbon construction innovation) and F4 (policy compliance), which are crucial for translating policies into action. The results layer focuses on the final outcome, represented by F12 (low-carbon construction waste recycling), which reflects the ultimate result of the system’s actions. The analysis focuses on the degree of impact (representing influence), the cause degree (representing causal factors), and stability (representing the system’s ability to resist changes), creating an integrated evaluation framework of “dynamics-attributes-structure”.

As shown in Table 15, the sensitivity analysis results indicate that the core structure of the system remains stable. The impact degree of F13 in the root cause layer changed by only ±0.3% (2.155 → 2.142), and the cause degree fluctuated by just ±0.7% (1.993 → 1.968), with changes less than 0.03. F15 showed complete stability, with the cause degree remaining constant at 1.159. In the intermediary layer, F9 experienced a ±18.8% fluctuation in cause degree, but the actual change was just ±0.03 (0.18 → 0.15), and its technical function remained stable. F4 showed a ±6.5% fluctuation in both adjusted scenarios but remained within acceptable limits. In the results layer, the impact degree of F12 fluctuated by ±0.8%, which is expected for final outcome factors. Overall, 85% of the factors retained their original hierarchical positions, and the average fluctuation in key indicators was controlled within ±0.5% for impact degree and ±4.1% for cause degree, both well below the industrial stability threshold of ±5%.

This analysis confirms that expert weight adjustments do not significantly alter the key structure of the model, which remains driven by the flow of “policy → technology → outcome”. Differences in experts’ backgrounds primarily caused small fluctuations in the intermediary links (F4/F9), and these fluctuations remained within acceptable engineering limits. This demonstrates that the model is resilient and stable, even when different expert opinions are considered. The model results provide valuable insights for decision-makers on low-carbon transition strategies, offering a reliable tool for policymakers and companies to develop tailored carbon-neutral strategies.

4. Discussion

4.1. Causal Analysis

As shown in Table 6, the top five influencing factors include low-carbon policy and regulatory guidance (F13), low-carbon consumer preferences (F15), low-carbon advocacy by social organizations and the media (F14), low-carbon construction technology innovation (F9), and low-carbon construction and operation concepts (F6). These factors play a critical role in the low-carbon building supply chain and are instrumental in driving its decarbonization. Specifically, low-carbon policy and regulatory guidance are crucial in promoting the industry’s overall low-carbon transition, as policy support provides the necessary institutional foundation for decarbonization. In parallel, low-carbon consumer preferences and advocacy efforts by social organizations and the media facilitate the promotion of low-carbon buildings through shifts in market and social environments. Furthermore, on the technological front, low-carbon construction technology innovations and low-carbon construction and operation concepts provide the technical and managerial support required for the effective implementation of decarbonization. Consequently, stakeholders in the building supply chain must focus on these key factors, develop appropriate strategies, and drive the low-carbon transition to ensure the successful achievement of low-carbon building goals.

As shown in Table 6, the factors most affected by external changes include low-carbon energy and material supply (F1), low-carbon supplier and contractor selection (F4), low-carbon property management (F5), low-carbon construction technology innovation (F9), and low-carbon construction waste recycling (F12). This indicates that these factors are highly sensitive to changes in the external environment, making them more susceptible to external influences, which directly impacts the implementation of the low-carbon building supply chain. Therefore, managers should pay particular attention to these factors, ensuring effective adaptation to external changes, thereby maintaining the stability of the supply chain and driving the achievement of low-carbon objectives. Notably, F9 (low-carbon construction technology innovation) is a special factor, as it is both highly influential and susceptible to external changes, playing a crucial role in the low-carbon transformation of the building supply chain.

As shown in the causality ranking in Table 6, factors such as low-carbon policy and regulatory guidance (F13), low-carbon consumer preferences (F15), low-carbon advocacy by social organizations and media (F14), low-carbon information sharing (F17), and low-carbon construction and operation concepts (F6) show high causality values. This indicates that these factors play a dominant role in driving the low-carbon transformation of the building supply chain and are less influenced by other factors. Specifically, low-carbon policy and regulatory guidance (F13) are key factors in driving the low-carbon transition, as the improvement and implementation of policies and regulations effectively guide the development of other factors. Low-carbon consumer preferences (F15) and low-carbon advocacy by social organizations and media (F14) promote the market and societal low-carbon transition by raising public awareness and demand. Low-carbon information sharing (F17) and low-carbon construction and operation concepts (F6) offer support and guidance for low-carbon building projects, ensuring that low-carbon goals are met throughout the design, construction, and operation phases. In summary, these factors play a proactive role in advancing the low-carbon transformation of the building supply chain and should be prioritized in management to ensure the successful achievement of low-carbon goals at all stages.

The centrality values in Table 6 show that each influencing factor plays a crucial role in the decarbonization of the building supply chain. The centrality ranking shows that F9 (low-carbon construction technology innovation), F4 (low-carbon supplier and contractor selection), and F6 (low-carbon construction and operation concepts) are critical core factors in the low-carbon building supply chain transformation. These factors play significant roles in the supply chain and can greatly impact the entire system’s operation. Notably, F9 (low-carbon construction technology innovation) is a key technical support for achieving efficient low-carbon building projects, highlighting the indispensability of technological innovation in the building supply chain. Furthermore, the high centrality and causality of F6 (low-carbon construction and operation concepts) indicate that it has a powerful influence and plays a pivotal role.

4.2. Hierarchical Analysis

According to the logical relationships between the system elements and the extraction results, this study presents the confrontation multilevel recursive order structure model, as shown in Figure 8. The model categorizes the influencing factors into ‘upward’ and ‘downward’ types, which have an antagonistic relationship with each other. In this paper, directed line segments represent the reachability relationships between influencing factors in the decarbonization of the building supply chain.

As shown in the figure, the bottom factor is the most fundamental one, representing an essential causal factor located at the lowest level of the hierarchical structure. The topology diagram shows that the bottom factors only send outward-directed line segments and do not receive influence from other factors. This indicates that these factors affect other influencing factors but are not directly influenced by other elements in the system. The bottom layer factors affecting the decarbonization of the building supply chain include {F13, F15}∪{F13, F15}, i.e., F13 (low-carbon policy and regulatory guidance) and F15 (low-carbon consumer preferences), which serve as the foundation for all other decarbonization measures, driving technological innovation, market demand, and system change.

Surface factors are the most direct influencing factors, belonging to the neighborhood causal factors and located at the top of the hierarchy. The topology diagram shows that the surface factors only receive directed segments from other factors and do not send out directed segments. According to the model results, the direct influencing factors for the decarbonization of the construction supply chain include {F1, F2, F3, F5, F7, F10, F12}∪{F1, F12}, i.e., F1 (low-carbon energy and material supply), F2 (low-carbon equipment supply), F3 (low-carbon transportation), F5 (low-carbon property management), F7 (low-carbon structural design), F10 (low-carbon construction process), and F12 (low-carbon construction waste recycling). These factors directly contribute to the decarbonization transformation of the construction supply chain by providing essential resources, equipment, technology, and management measures.

In addition to the bottom and surface factors, other factors are located in the second and third tiers, referred to as mid-tier factors, which play a mediating role in the overall framework. Therefore, managers should pay special attention to these mid-tier factors and establish an effective monitoring and evaluation mechanism to regularly track and assess the impact of transition factors on the decarbonization of the construction supply chain. By collecting and analyzing relevant data, managers can gain a comprehensive understanding of the trends and impacts of transition factors, allowing them to adjust strategies promptly to minimize potential adverse impacts and ensure the successful achievement of low-carbon goals.

4.3. Comprehensive Analysis

Through the Fuzzy-DEMATEL-AISM analysis, this study reveals the critical role of low-carbon policy and regulatory guidance, low-carbon advocacy by social organizations and media, and low-carbon consumer preferences in the decarbonization of the building supply chain, consistent with existing research. For instance, Tang D’s bibliometric analysis indicates that government actions, particularly policies and regulations, significantly promote the low-carbon transition of the building supply chain. The study found that government actions play a key role in reducing carbon emissions during the material production, construction, and operation stages, especially in green supply chain management, green building decisions, and energy-saving behaviors. This aligns with the role of low-carbon policy and regulatory guidance, demonstrating that policies and regulations are a core driver of technological innovation and market demand [46].

Muok B O investigated the role of civil society organizations in low-carbon innovation, particularly in Kenya. The study found that civil society plays a vital role in raising public awareness and promoting low-carbon behaviors, particularly in grassroots-level low-carbon transitions. This is consistent with the role of low-carbon advocacy by social organizations and media, highlighting the key role of social organizations in promoting low-carbon innovation [47].

Additionally, S Du utilized emission-sensitive demand and cost functions to study the impact of consumer low-carbon preferences on supply chain emission reduction strategies. The research found that low-carbon consumer preferences can simultaneously increase channel profits and emission reduction effectiveness under certain conditions, thereby promoting the low-carbon transformation of the building supply chain [48]. The AISM analysis in this study further corroborates these findings and provides a new perspective, highlighting the importance of low-carbon construction technology innovations in the decarbonization process of the building supply chain, offering new insights and references for future research.

4.4. Responses and Recommendations

A three-tier strategy can be proposed to promote the decarbonization of the construction supply chain, based on the interrelationships among the 17 influencing factors (Figure 9).

At the foundational level of the low-carbon building supply chain transformation, the government plays a central role in driving the transition. By formulating and implementing low-carbon policies and regulations, the government provides support for various low-carbon actions, facilitating the achievement of low-carbon objectives. The government should introduce targeted low-carbon policies, establish clear low-carbon building standards, and provide support through fiscal and tax incentives. These policies should cover all stages of building design, construction, and operation, ensuring that the market and industry can effectively adapt according to policy guidance. For instance, the European Green Deal is a successful case that demonstrates the crucial role of policy in driving the low-carbon transformation of the construction industry. The goal of the European Green Deal is to achieve carbon neutrality by 2050, with the building sector, as a major energy-consuming field, carrying a significant responsibility. According to a report from the European Commission, by 2030, the energy efficiency of buildings must be improved by at least 32.5%, and carbon emissions from the building sector must be reduced by at least 40% [49]. This policy measure has promoted the application of green building technologies and materials, as well as the energy-saving renovation of building projects.

In addition, the government should further promote the low-carbon transition by fostering low-carbon consumer preferences. By increasing public awareness of low-carbon buildings, the government can accelerate the adoption of low-carbon consumption. Environmental advocacy, green certification labels, and other means help raise consumer awareness and preference for low-carbon building products, thus stimulating market demand. Increasing support for low-carbon building products will enable consumer demand for green, low-carbon products to become a driving force for industry transformation.

The second level emphasizes the role of enterprises in the low-carbon transition. As the direct implementers of low-carbon measures, enterprises drive the transition through technological innovation, management optimization, and supply chain management. First, promoting low-carbon technology innovation and application is a key task for enterprises. They should strengthen the research, development, and application of low-carbon building technologies, adopting advanced low-carbon construction techniques and energy-saving technologies to reduce energy consumption and carbon emissions in building projects. For instance, the BedZED project in London, UK, is a typical successful case. This project is the world’s first zero-energy residential area, employing low-carbon technologies such as solar panels, rainwater harvesting systems, and high-efficiency insulation materials. According to project data, BedZED’s annual energy demand is reduced by 81%, and carbon emissions are reduced by 56% compared to conventional buildings. Furthermore, the project effectively improves resource usage efficiency and reduces environmental impact. This demonstrates that by adopting advanced low-carbon building technologies and green construction techniques, enterprises can significantly reduce carbon emissions in building projects and enhance energy efficiency, thereby driving the implementation of the low-carbon transition [50].

The third tier should focus on the project level, as specific implementation measures directly determine the achievement of low-carbon building goals. Firstly, projects should emphasize low-carbon structural design from the design phase, selecting environmentally friendly building materials and incorporating low-carbon optimization designs. For example, the Shanghai Tower in China adopted a passive design concept, optimizing natural lighting, ventilation, and insulation, thus reducing building energy consumption. Project data shows that this building is 50% more energy-efficient than traditional skyscrapers, maintaining a low-energy state in the long term [51]. Secondly, promoting low-carbon energy design is crucial. Projects should reduce energy consumption during operation by integrating renewable energy systems such as solar and wind energy, while also implementing smart energy management systems to optimize energy usage in real time. Finally, implementing low-carbon building waste recycling is equally important. For instance, the Central Park project in Sydney implemented a zero-waste policy, recycling and reusing approximately 90% of construction waste, significantly reducing carbon emissions and promoting the circular use of building resources. Through these measures, projects can effectively support the achievement of low-carbon goals [52].

4.5. Policy and Practical Implications

Based on the key driving factors (such as policies and technologies) and the system hierarchy identified in the study, the following actionable recommendations are proposed. At the policy level, the government should lead institutional innovation to promote low-carbon transformation. First, a combination of carbon taxes and subsidies can effectively guide construction companies to adopt low-carbon measures. By imposing a tiered carbon tax based on emissions and offering subsidies to companies implementing low-carbon technologies, the government can incentivize companies to reduce their carbon footprint. For instance, the European Union provides a 20% tax reduction for companies using recycled building materials. Additionally, mandatory carbon disclosure is crucial. Developers should be required to disclose a building’s carbon emissions throughout its lifecycle at the time of sale, similar to the way nutrition labels are used on food products. This not only helps consumers make low-carbon choices but also stimulates market demand, positively influencing F15.

At the corporate level, construction companies should form technology alliances to jointly promote the research and application of low-carbon technologies. Large construction firms can establish a “Low-Carbon Technology Alliance” to share low-carbon technology patents (such as the prefabricated assembly technology in F9), thereby reducing research and development costs. For instance, Toyota in Japan collaborated with construction companies to develop low-carbon modular housing, resulting in a 30% cost reduction. Moreover, companies should enhance supply chain management by implementing green certifications, evaluating suppliers’ low-carbon performance (such as F4’s low-carbon material supply), prioritizing collaboration with highly rated suppliers, and incorporating this criterion into contractual agreements. This supply chain collaboration will contribute to improving the low-carbon performance of the entire industry.

At the project level, practical environmental measures must be implemented. One key measure is recycling construction waste, where smart recycling stations (corresponding to F12) can be installed at project sites to automatically sort waste materials such as steel bars and concrete, aiming for a recycling rate of at least 85%. For instance, a construction site in Shanghai reduced landfill waste by 60% using this method. Additionally, energy consumption monitoring should be enhanced by using smart meters to track the building’s energy usage (corresponding to F5 property management), with the data accessible to property owners via an app. A commercial building in Shenzhen saved 2 million RMB in annual electricity costs by implementing this measure. These practical measures can effectively drive the development of low-carbon buildings and lay a solid foundation for achieving sustainable construction goals.

4.6. Core Contributions

Based on the discussion above, this study systematically identifies 17 key influencing factors for decarbonization in the building supply chain, considering the interrelationships among government, consumers, and construction supply chain enterprises from an upstream-to-downstream perspective. This multi-dimensional and systematic analysis addresses the gap in the existing literature, particularly in the area of how multiple factors work synergistically to drive low-carbon transformation, an aspect that has been less explored in previous research. Furthermore, this study is the first to apply the Fuzzy-DEMATEL-AISM method to the decarbonization of the building supply chain. Unlike previous studies, the method used in this paper comprehensively examines the dynamic relationships among the influencing factors, effectively minimizing subjective bias. Therefore, this research not only broadens the scope of low-carbon building supply chain studies but also fills the academic gap in the application of methodologies.

Practically, the policy-enterprise-project action plan provides governments, firms, and project teams with step-by-step strategies—from carbon taxes to waste recycling—to operationalize findings. These contributions collectively bridge academic insights and real-world decarbonization challenges in the construction sector.

5. Conclusions

This study employs bibliometric analysis to identify theoretical research gaps in the construction supply chain field. From an upstream-to-downstream perspective, and considering the interrelationships among the government, consumers, and construction supply chain enterprises, it systematically and comprehensively identifies 17 key influencing factors for the decarbonization of the construction supply chain. This study addresses a gap in the existing literature, which lacks an integrated approach to understanding how multiple factors work synergistically to drive the low-carbon transition of the construction supply chain. To minimize subjective bias, this study is the first to apply the Fuzzy-DEMATEL and AISM methods to the decarbonization of the construction supply chain, exploring the complex relationships, classifications, and relative importance of these factors. Subsequently, a Recursive Structural Model (RSM) is developed to systematically describe the key factors and their interactions in the decarbonization process, thus overcoming the limitations of previous studies that relied solely on single analytical methods to examine influencing factors.

The findings emphasize the critical roles of low-carbon policy guidance and low-carbon construction technology innovation in the decarbonization of the construction supply chain. Furthermore, the results indicate that an effective low-carbon transition requires coordination among government policies, corporate actions, and project-level implementation, particularly in facilitating the adoption of low-carbon technologies and optimizing resource allocation. Based on these insights, this study proposes a three-tier implementation framework: policy guidance → corporate innovation → project execution. Specifically, governments should steer market transformation through carbon taxation and transparency mechanisms (e.g., mandatory carbon disclosure), enterprises should collaborate on low-carbon technology R&D (e.g., modular construction alliances), and project teams must enforce waste recycling and energy monitoring (e.g., ≥85% on-site recycling targets). This integrated framework systematically advances the low-carbon transition of the construction supply chain. These findings directly address the global decarbonization challenges outlined at the beginning of the paper, particularly the need to reduce emissions within the construction supply chain and overcome fragmented improvement measures.

However, this study has certain limitations. First, the limited involvement of industry practitioners may have affected the depth of understanding of real-world challenges and the feasibility of practical applications of decarbonization strategies. Second, while the Fuzzy-DEMATEL-AISM method provides valuable insights, it does not fully capture the dynamic interactions among the identified factors. In particular, the complexity and evolution of these interactions may not be adequately reflected in rapidly changing market and policy environments.

To address these limitations, an interesting direction for future research is the integration of the fuzzy-DEMATEL method with machine learning models. Machine learning techniques, such as regression analysis, neural networks, and deep learning, can enhance the fuzzy-DEMATEL approach, particularly in analyzing larger and more complex systems. This integration could provide deeper insights into the dynamic and nonlinear relationships between decarbonization factors, thereby improving the model’s predictive capabilities. For instance, Hassan [53] investigated the combination of fuzzy-DEMATEL with federated learning to assess data security in autonomous network management. Their study demonstrated that this integration enhances data security, particularly when dealing with sensitive information. By leveraging federated learning, private data can remain protected while enabling the integration of heterogeneous datasets to optimize system decision-making and predictive accuracy.

These advancements will contribute to a more comprehensive understanding of decarbonization pathways and strengthen the construction industry’s capacity for sustainable transformation.