A System-Based Framework for Reducing the Digital Divide in Critical Mineral Supply Chains

Abstract

The widening digital divide within the Global Critical Mineral Resource Supply Chain (GCMRS) 4.0 creates significant barriers to cross-border governance and operational efficiency. To quantify and address this disparity, this study identifies 20 Critical Success Factors (CSFs) through expert interviews with 15 industry specialists in South Korea. A hybrid multi-criteria decision-making framework integrating Fuzzy DEMATEL, Analytic Network Process (ANP), and the Choquet integral is developed to map causal relationships and determine factor weights. The empirical results reveal a distinct ‘technology-first’ dependency. Specifically, Scalable Technical Solutions and Cloud Computing Access emerge as the primary driving forces with the highest global weights, while Digital Investment Subsidies serve as the central hub for resource allocation. Unlike generic governance models, this study provides a quantifiable decision-making basis for policymakers. It demonstrates that bridging the hard infrastructure gap is a prerequisite for the effectiveness of soft collaborative mechanisms in the critical mineral sector.

1. Introduction

Technological advancements and the clean energy transition have rendered the global supply chain for critical mineral resources (GCMRS) increasingly indispensable [1,2]. Effectively managing these resources is vital for international economic stability and national strategic security [3]. Currently, the emergence of Industry 4.0 defines the essence of GCMRS 4.0, which aims to build a digitally integrated, transparent, and resilient network [4,5]. However, a persistent technological disparity significantly obstructs this vision. Although developed countries have accelerated modernization [6], this progress has inadvertently widened the gap with infrastructure-limited regions. This growing disparity challenges the overall effectiveness and inclusiveness of the global supply system [7,8].

This inequality is particularly severe in resource-rich but technologically underdeveloped nations [9,10]. While scholars originally defined the digital divide as disparities in internet access [11], it has evolved into a structural bottleneck across the entire value chain [12]. A critical manifestation of this issue is the ‘missing middle’ phenomenon. Specifically, while innovative startups embrace digitalization, traditional small and medium-sized enterprises (SMEs) continue to lag due to resource constraints and limited managerial capabilities. This creates a fragmented landscape where the digital transformation efforts of key production partners remain stagnant [13].

Consequently, digital inequality directly impairs the ability of partners to acquire information and achieve operational efficiency [14]. It amplifies information asymmetries in governance and weakens the long-term sustainability of the supply chain [6]. Organizations that fail to implement timely transformation risk obsolescence in shifting market conditions [4]. Given that this divide results in low overall efficiency and weak coordination, it is imperative to identify the Critical Success Factors (CSFs) that can bridge this gap and foster equitable development [15].

Despite growing attention to digitalization in the GCMRS, scholars continue to debate its implications. Some studies emphasize the potential of technologies such as blockchain, the Internet of Things (IoT), and digital twins to improve supply chain transparency. These technologies allow stakeholders to trace minerals throughout the entire process—from extraction to transportation—reduce illegal mining, and optimize resource allocation [16,17]. However, other studies highlight that poor infrastructure in many regions prevents the widespread implementation of such technologies [4].

Some scholars also focus on enhancing the operational efficiency of mineral supply chains through digital tools. For instance, demand forecasting and real-time monitoring technologies help allocate resources effectively and adjust mineral transportation dynamically [17,18,19]. However, few explicitly investigate how these technologies can bridge the digital divide. Meanwhile, a growing body of research examines how digital technologies contribute to sustainability and risk management in mineral supply chains [3,20]. Although these studies offer valuable insights into building greener and more resilient supply chains, most focus primarily on digitally advanced economies. As a result, the literature gives insufficient attention to the balanced deployment of digital technologies across the GCMRS, especially among diverse international partners. A substantial gap remains in addressing digital inequality in this global context.

To address this gap, this study develops a comprehensive strategic framework to reduce the digital divide in critical mineral supply chains 4.0. It poses the following research questions (RQs):

RQ1: What are the CSFs for reducing the digital divide in the GCMRS?

RQ2: What are the distinctions and interrelationships among these CSFs?

To address these questions and bridge existing research gaps, this study constructs a strategic framework for the GCMRS that is theoretically anchored in three complementary perspectives: Innovation Diffusion Theory, the Resource-Based View (RBV), and Social Network Theory. By integrating these lenses, the study provides a structured approach that explicitly aligns digital transformation goals with organizational capabilities and network dynamics.

Innovation diffusion theory underscores the importance of widespread adoption of accessible and efficient digital technologies within the GCMRS [21]. RBV posits that firms can gain competitive advantages by acquiring and managing rare, valuable, and inimitable technical resources [22]. Based on this perspective, companies must prioritize the development of digital technologies and professional competencies—especially in underdeveloped regions—to help partners strengthen their technical capacity and integrate resources across the supply chain 4.0 [19]. Social network theory highlights the role of network participants and their interactions in shaping collaborative outcomes [23]. By fostering collaboration through strong and weak ties, organizations can share technical and knowledge resources, thereby enhancing their collective innovation capacity [20].

Grounded in these theoretical foundations, this study proposes system-oriented strategies from the literature to promote equitable access to and efficient use of digital technologies. Ultimately, these strategies aim to advance digital equality across the GCMRS.

By identifying the key drivers of the digital divide, this study constructs a governance framework for the GCMRS 4.0. The remainder of the paper is structured as follows: Section 2 reviews the literature on the digital divide and identifies relevant CSFs. Section 3 outlines the research methodology. Section 4 presents the results of the data analysis. Section 5 discusses the research findings in depth. Section 6 concludes by highlighting the study’s theoretical and managerial contributions and addressing its limitations.

2. Literature Review

2.1. Digitalization of the Mineral Resource Supply Chain

In recent years, scholars have extensively researched the digitalization of mineral resource supply chains. This body of work falls into three main areas: the role of digital technologies in enhancing the sustainability of mineral resource supply chains [7], the application of digital technologies to improve supply chain efficiency, and the use of digital tools in decision-making and risk management [19].

Studies on emerging digital tools, most notably the Internet of Things (IoT) and blockchain, show that these technologies play a pivotal role in enhancing supply chain traceability, thereby supporting sustainability goals, reducing illegal mining, and minimizing environmental impacts. In particular, blockchain technology records information across all stages—from mineral extraction to product distribution—thereby promoting accountability, compliance, transparency and collaborative efficiency [24]. Integrating ecological innovation with digital technologies optimizes resource utilization and provides technical support for achieving sustainable development goals [20].

However, some studies highlight that developing countries often face challenges such as inadequate infrastructure and disparities in technical capacity, which contribute to a persistent digital divide that hampers the sustainable development of mineral supply chains [25]. In response, other studies emphasize that policy support and international collaboration, particularly through the promotion of ecological innovation and energy transition technologies, can overcome barriers such as the digital divide, resource management imbalances, and limited technology diffusion [20]. Collectively, these findings underscore that digital technologies hold significant potential for enhancing sustainability in mineral resource supply chains. However, their widespread adoption depends on stronger policy support and sustained investment in digital infrastructure and capacity-building.

Research on improving efficiency in mineral resource supply chains has focused on applying data-driven predictive models and intelligent optimization algorithms [4]. Studies demonstrate that machine learning technologies can forecast demand, optimize resource allocation, and thereby enhance supply chain responsiveness and reduce operational costs [26,27]. The integration of real-time data analysis with sensor networks enables dynamic adjustments to mineral transportation systems, improving logistics efficiency [18]. Some studies have explored the combination of digital twin technology and blockchain, finding that it improves transparency and operational performance in mineral resource supply chains [17].

Despite these advances, underdeveloped infrastructure in many developing regions remains a significant barrier to improving supply chain efficiency [28]. To address this, studies suggest adopting intelligent management tools alongside supportive policy measures to increase technology uptake and infrastructure investment, thereby enhancing operational efficiency [29]. Consequently, the literature suggests that digital technologies function not merely as support tools but as structural prerequisites for operational efficiency in mineral resource supply chains.

The application of digital technologies in decision-making and risk management within mineral resource supply chains has also attracted considerable scholarly attention [19,30]. Researchers have developed multi-criteria decision analysis frameworks to identify potential risk factors and enhance resilience to disruptions [31]. These studies identify key factors influencing the adoption of digital technologies for risk management and propose solutions involving technological integration and supportive policies [3].

However, scholars also point out that policy and economic barriers in developing countries hinder technology implementation [32]. They argue that international cooperation and resource sharing are necessary to address the complexity and high concentration of risks in mineral resource supply chains [33]. Other research highlights blockchain’s role in enhancing transparency and responsiveness to uncertainty, while artificial intelligence significantly enhances scientific and reliable risk prediction and management [28]. Collectively, these studies demonstrate that the deep integration of digital technologies supports scientific decision-making and risk management while improving the resilience of the global mineral resource supply chain [20].

2.2. Theoretical Development and Critical Success Factors

Establishing a robust theoretical foundation is essential for analyzing the digital divide in the GCMRS. It allows for a systematic examination of technological adoption barriers and supply chain management complexities. This section discusses the critical success factors (CSFs) for managing the digital divide in the context of the critical mineral supply chain through three core theoretical perspectives: innovation diffusion theory, the resource-based view (RBV), and social network theory [34].

The theory of innovation diffusion provides a framework for examining how social systems adopt technologies and ideas [21]. It explores adoption pathways, influencing factors, and conditions that facilitate application [21]. Scholars have widely applied this theory in agriculture, healthcare, and information technology, making it highly relevant to technology adoption in supply chain management [35]. In supply chains, the theory highlights multi-level dynamics, emphasizing relative advantage, compatibility, observability, and the influence of socio-cultural environments and organizational characteristics on adoption [35]. It posits that enhancing trialability and visibility can strengthen trust among partners and accelerate adoption [21].

Within mineral resource supply chains, innovation diffusion theory provides an essential theoretical basis for addressing the digital divide. CSFs include the deployment of efficient and widely accessible technological solutions—such as cloud computing and mobile technologies—that can reduce access barriers in developing regions [21]. Demonstration projects, which pilot new technologies on a small scale, play a vital role in showcasing tangible benefits, validating feasibility, and providing data to support large-scale adoption [36]. Accelerating the diffusion of digital technologies in the Global Mineral Resource Supply Chain requires enhancing the perceived relative advantage and compatibility while strategically using demonstration projects to reduce uncertainty and bridge the digital divide [18].

RBV posits that a firm’s competitive advantage originates from strategic resources that are scarce, inimitable, and irreplaceable, including tangible assets (such as equipment and infrastructure), intangible assets (such as brand equity and intellectual property), and human capital (such as technical skills and expertise) [37]. To realize full strategic value, firms must integrate and manage resources effectively [38]. In dynamic and rapidly evolving market environments, RBV offers an analytical lens for optimizing resource allocation and strengthening resilience and competitiveness [37].

Various industries have applied this framework extensively, including manufacturing, service industries, and supply chain management, offering theoretical support for improving collaboration and resource-sharing capabilities across supply networks [37]. In mineral resource supply chains, RBV emphasizes enhancing partners’ digital capabilities through skill development and resource sharing to narrow capability gaps and reduce information asymmetries [4]. For instance, continuous learning platforms can equip partners with digital tools and practical knowledge, improving technical and managerial competencies [19]. Leadership development also plays a key role in enabling supply chain leaders to adapt to technological change [39]. Cross-organizational resource sharing fosters trust and enables localized technical support, contributing to more balanced development across the supply chain [20]. Accordingly, RBV support strategies that promote digital skill development, resource sharing, and leadership capacity to close the digital divide [32].

Social network theory focuses on how interactions, information flow, and resource sharing among supply chain participants facilitate technology diffusion and enhance innovation capabilities [4]. It views each partner as a node, with the network’s density, centrality, and tie strength shaping information and resource dissemination [40]. In mineral resource supply chains, effective social networks depend on information transparency and interaction among partners, particularly in contexts with information asymmetries and disparities in digital capability.

Key CSFs include implementing information transparency policies that ensure timely access to accurate data, thereby fostering trust [18]. Encouraging active decision-making and network activities also helps distribute knowledge and resources equitably—a critical strategy for overcoming digital divides in developing regions [20].

In summary, these three theoretical frameworks provide a multifaceted foundation for understanding and addressing the digital divide in the Global Mineral Resource Supply Chain. They highlight the importance of technology deployment, skill development, collaboration, and supportive governance mechanisms [41]. Drawing on these insights, this study identifies three key strategies to reduce the digital divide: technology deployment, skill development and resource sharing, and policy and regulatory support.

First, technology deployment should ensure that digital tools offer relative advantages, compatibility, and usability, enabling equitable access to technologies like cloud computing and mobile platforms [21]. Second, skill development and resource sharing help build partners’ technological application capacity through continuous training and shared resources, thereby enhancing digital competencies and management capabilities [42].

Finally, policy support and network collaboration are also vital. Policies that promote information transparency and cross-organizational collaboration can foster trust and cooperation, thereby accelerating digital transformation [20]. Collectively, these strategies provide a pathway to close the digital divide, improve efficiency, and strengthen the GCMRS’s capacity for sustainable development.

3. Proposed Framework

3.1. Analysis Methodology Process

We structured this study’s methodological framework into three main stages. First, we constructed the indicator system which serves as the analytical cornerstone. This construction was scientifically justified by grounding the selection process in the theoretical integration of innovation diffusion theory, the resource-based view, and social network theory. To ensure rigor, we initially identified a broad pool of potential factors from authoritative databases such as Web of Science and Scopus. Subsequently, these factors underwent a rigorous screening process involving expert consultation to verify their content validity and operational applicability. This process ultimately yielded 20 Critical Success Factors (CSFs) as detailed in

Table 1. Definition of critical success factors.

No.	Critical Success Factor	Definition	Sources
1	Scalable technical solutions	Provide scalable technology suitable for different enterprise sizes and capabilities to ensure that all partners can participate equally in supply chain activities. Reduce technical barriers for small partners through technological flexibility.	[43]
2	Cloud Computing Access	Leveraging cloud technology to reduce partners’ need for expensive infrastructure investments, providing low-cost, high-efficiency solutions for businesses in developing regions, while promoting centralized data management and rapid deployment.	[43,44]
3	Mobile Technology Integration	By leveraging mobile technology, which developing countries widely use, partners in remote areas can also access global supply chain systems, thereby addressing the digital divide in areas with weak infrastructure.	[12]
4	Regional Connectivity and Technology Solutions	Provide regionalized digital connectivity and technical solutions for resource-constrained areas, reducing the constraints of the digital divide on supply chain digitization through low-cost and efficient means.	[45]
5	Interoperable System	Ensure that new technology tools are compatible with existing systems to promote seamless integration and data flow within the supply chain, reducing coordination costs caused by technical differences.	[46,47]
6	Blockchain technology for supply chain transparency	Utilizing blockchain technology to achieve full traceability of minerals from extraction to transportation, enhancing supply chain transparency and compliance, effectively reducing illegal activities, and optimizing resource allocation.	[48,49]
7	Application of Digital Twin Technology	Create virtual supply chain models using digital twin technology to monitor and optimize logistics networks in real time, improve resource management efficiency, and enhance the supply chain’s responsiveness to external changes.	[50,51,52]
8	Supply Chain Environmental Data Monitoring Tool	Development tools monitor environmental data (such as carbon emissions and resource usage) in the supply chain, supporting the integration of green and digital supply chains.	[53,54]
9	Digital Literacy Program and Learning Platform	Implement a digital literacy enhancement program through a digital learning platform to provide supply chain partners with the latest digital tools and skills training, thereby enhancing their digital capabilities within the supply chain.	[15,43]
10	Digital Transformation Leadership Training	Provide supply chain managers with leadership skills to respond to digital transformation, thereby better driving enterprise digital transformation and achieving efficient supply chain operations.	[55]
11	Supply chain partner technical support and customized capacity building	Provide customized technical support and capacity-building services, design personalized solutions based on the specific needs of partners, and help them improve their digital technology application capabilities.	[34,44]
12	Skill upgrade incentives	Through incentive mechanisms, encourage employees and partners to participate actively in learning and practicing digital technologies, promoting enthusiasm for technological innovation and skill upgrades.	[39]
13	Building a Diversified Supplier Network	By establishing a diversified supplier system, we enhance the resilience of the supply chain in the face of unexpected circumstances, ensuring its continuity and flexibility.	[33,45]
14	Resource integration and knowledge sharing platform	Establish a shared platform to improve overall supply chain collaboration through centralized management of knowledge, technology, and resources, promoting close connections between supply chain nodes and resource optimization.	[47,56]
15	Localized Content	Provide digital technology content that is tailored to local cultures and languages to enhance partner acceptance and usage effectiveness, particularly in diverse scenarios within multinational supply chains.	[12,43]
16	Data protection regulations	Establish robust data protection regulations to enhance partners’ trust and confidence in digital technologies, ensuring the secure flow and transparency of supply chain data.	[40,57]
17	Multi-level digital standardization coordination	Combining internal supply chain and international standardization requirements, promote multi-level digital standardization cooperation to improve data flow and system integration efficiency.	[46,47]
18	Digital Investment Subsidy	Provide financial support to smaller supply chain partners to help them invest in the necessary digital infrastructure, reduce the digital divide, and promote digital transformation.	[45,58]
19	Real-time feedback mechanism for the supply chain	Through a real-time feedback system, obtain operational improvement suggestions from supply chain partners, optimize overall operational efficiency, and ensure the timeliness and accuracy of supply chain management decisions.	[55,59]
20	Collaborative Work Environment Platform	Establish a digital collaboration platform to promote efficient communication and resource sharing among upstream and downstream enterprises, and strengthen mutual trust and collaboration capabilities within the supply chain.	[43,53,56]

.

Second, we applied the Fuzzy Decision-Making Trial and Evaluation Laboratory (FDEMATEL) method to analyze the causal relationships among these factors. This method was explicitly selected for its ability to handle the ambiguity and fuzziness inherent in expert judgments, providing a more robust interaction analysis than traditional crisp value approaches [60]. Furthermore, to explore the hierarchical structure and interdependencies among the CSFs, we integrated the Analytic Network Process (ANP) method into the FDEMATEL framework. Unlike the Analytic Hierarchy Process (AHP) which assumes independence among factors, ANP is specifically designed to model the feedback mechanisms and complex interrelationships characteristic of supply chain systems [61].

Finally, we employed the Choquet integral to conduct an aggregated assessment of the CSFs. This technique was chosen to address the non-additive nature of the criteria, allowing the model to capture the synergistic effects among interdependent factors that linear weighting methods often overlook [62]. This rigorously constructed framework offers theoretical guidance and decision-making support for governance practices across all stakeholders.

Finally, we used the Choquet integral to conduct an aggregated assessment of the CSFs, thereby constructing a scientific and systematic optimal factor framework. This framework offers theoretical guidance and decision-making support for governance practices across all stakeholders.

During the empirical research phase, we first identified government agencies, mining companies, industry associations, and digital platform operators involved in South Korea’s key mineral resource supply chains as target subjects. We selected representative organizations using publicly available information and prevailing industry trends, then contacted key personnel from these organizations to explain the research objectives and significance, inviting them to participate.

Over the next four weeks, the team received 31 responses via telephone and email. Although some experts could not participate in in-depth interviews due to scheduling conflicts, 15 experts (see

Table 2. Expert Demographics.

Profile Information	Number	Percentage (%)
Organization Type
Government	3	20.00%
Academic Institution	4	26.70%
Enterprise	6	40.00%
Association	2	13.30%
Staff Size
<50	1	6.70%
51–200	3	20.00%
201–300	1	6.70%
>301	10	66.70%
Experience
<5	1	6.70%
5–10	5	33.30%
11–15	3	20.00%
>16	6	40.00%
Education Level
Postgraduate	9	60.00%
Bachelor’s Degree	6	40.00%
Position
Department manager and above	8	53.30%
Professor/Researcher	4	26.70%
Government Manager	3	20.00%

) with extensive research or practical experience in mineral resource supply chains and digital governance ultimately took part in the study. These experts understood the operational logic of cross-border resource trade systems and offered deep insights into how digital platforms enable supply chain collaboration.

To collect data, we combined webinars, remote interviews, and online questionnaires. We first provided a systematic overview of the study’s theoretical foundations and technical approach, then held structured discussions on factor identification and importance ranking. Subsequently, we distributed anonymous questionnaires to collect expert scoring results for the CSFs, thereby enhancing the objectivity and scientific rigor of the data.

Figure 1 illustrates this study’s research process.

3.2. Methodological Overview

3.2.1. Fuzzy-DEMATEL Overview

The Fuzzy DEMATEL method integrates fuzzy set theory, matrix modeling, and graph-theoretical analysis to uncover the causal relationship structure among factors within complex systems. It relies on expert judgment to construct a structural map of fuzzy causal relationships, thereby visually representing the interaction mechanisms within the system.

During implementation, we first applied the central point method (Formula (1)) to defuzzify the triangular fuzzy numbers derived from expert evaluations, thereby constructing the initial direct influence matrix. We then used Formula (2) to normalize this matrix, producing a standardized direct influence matrix that forms the foundation for subsequent causal strength analysis. Next, we applied Formula (3) to convert the normalized direct influence matrix into a total influence matrix, capturing the direct and indirect relationships among factors. This step provided a clearer understanding of the overall influence and dependency of each factor, supporting the visualization and identification of the causal structure.

$$Q={\displaystyle \frac{\left(r-m\right)+\left(n-m\right)}{3}}+m$$

(1)

$$T={\displaystyle \frac{A}{max\left[max\left(\sum _{j=1}^n g_{ij}\right),max\left(\sum _{i=1}^n g_{ij}\right)\right]}}$$

(2)

$$G=T(I-T{)}^{-1}$$

(3)

In this context, $T_{ij}$ indicates how strongly factor i influences factor $j$, and $i$ refers to the identity matrix. These interdependencies allow the computation of prominence (D + R), relation (D − R), and the influence degree of each factor, where D indicates the extent to which a factor influences others, and R reflects the extent to which it is influenced. Formulas (4)–(7) help us obtain these values.

$$D={\sum }_{j=1}^n T_{ij}$$

(4)

$$C={\sum }_{i=1}^n T_{ij}$$

(5)

$$E=D+C$$

(6)

$$R=D-C$$

(7)

3.2.2. Analytic Network Process Overview

In the ANP, the analysis of interdependencies among factors culminates in the derivation of the limit hypermatrix, G. We built the ANP network structure diagram by integrating the infinite hypermatrix generated from FDEMATEL, enabling a systematic analysis of influence relationships among indicators.

From the infinite supermatrix, we extracted the column vector $w$ representing the global weights. Then, we multiplied this vector element-wise by the centrality vector obtained from the DEMATEL analysis to compute each factor’s comprehensive weight. We express the limit supermatrix as $\underset{N\to \infty }{lim} \left({\displaystyle \frac{1}{N}}\sum _{k=1}^N W^k\right)$:

$$Z={\displaystyle \frac{MW}{\sum _{i=1}^n MW}}$$

(8)

3.2.3. Choquet Integral Overview

To compute the Choquet integral, we first derive the weight distribution of each key indicator from the limit supermatrix (i.e., the converged supermatrix) obtained via the ANP method. We then introduce a fuzzy measure to capture each indicator’s importance as perceived by experts. We obtain these fuzzy-measure values by having experts rank the relative priorities of the factors, thereby quantifying the subjective preference information associated with each indicator.

To address the challenge of data collection, we introduce the $\lambda$ − Additivity axiom. In a fuzzy metric space $(X,\beta ,g)$, if $\lambda \in \left(-0.5,0.5\right)$and $A\in \beta$, then $A\cap B=\varnothing$, then the fuzzy metric $g$ is $\lambda$ − Additivity (see Formula (9)). This type of fuzzy metric is the $\lambda$ − fuzzy metric because it must satisfy $\lambda$ − Additivity, commonly referred to as the Kanno metric.

$$g\left(A\cup B\right)=g\left(A\right)+g\left(B\right)+\lambda g\left(A\right)g\left(B\right),A\cap B=\varnothing$$

(9)

Finally, we obtain the fuzzy measure $g\left(X\right)$ by summing the values in each column of the DEMATEL matrix, as described in Formula (10).

$$gI\left(h,g\right)=\left\{\begin{array}{cc}{\sum }_{i=1}^n h_i·g_i,\hfill & \mathrm{when}\lambda =0\hfill \\ {\displaystyle \frac{1}{\lambda }}\left({\prod }_{i=1}^n \left(1+\lambda ·g_i\right)-1\right),\hfill & \mathrm{when}\lambda \ne 0\hfill \end{array}\right.$$

(10)

Here, $h_i$ denotes the weight assigned to the i-th indicator, while $g_i$ represents its corresponding fuzzy measure. According to the formula $g(A\cup B)=g\left(A\right)+g\left(B\right)+\lambda g\left(A\right)g\left(B\right)$, when $\lambda \ne 0,g\left(A\cup B\right)$ is non-additive, $\lambda =0,$ $g\left(A\cup B\right)$ becomes additive. That is, if there is a relationship between ${\lambda }_{ij}\ne 0,h_i$ and $g_j$, and if there is no relationship between ${\lambda }_{ij}=0h_i$ and $g_j$, in the fuzzy metric space $(X,\beta ,g)$, let $k$ be a measurable function mapping $X$ to $[-0.5,0.5]$. For simplicity, we call the fuzzy integral $\int \mathrm{h}\mathrm{d}\mathrm{g}$ the Choquet integral.

4. Results

This section applies a three-stage methodology that integrates Fuzzy DEMATEL, ANP, and the Choquet integral to analyze the CSFs for reducing the digital divide in the GCMRS. The step-by-step analytical process clarifies the relative importance of each factor and explains the mechanisms of their interactions.

4.1. Fuzzy DEMATEL Results

Table 1, Table 2,

Table 3. Direct influence matrix.

CSF	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17	C18	C19	C20
C1	0	0	4	3	1	1	4	0	3	0	0	3	4	3	4	3	4	3	0	4
C2	4	0	1	3	1	0	4	4	3	1	4	3	1	1	0	0	3	4	0	4
C3	0	3	0	4	1	3	0	4	3	1	0	4	1	4	3	4	4	0	1	1
C4	1	1	0	0	1	0	3	0	3	3	4	3	1	4	1	1	4	1	0	1
C5	3	3	3	3	0	4	0	1	0	3	1	1	0	3	3	4	1	0	4	3
C6	3	4	1	4	0	0	3	0	1	4	4	1	4	4	1	4	1	0	3	3
C7	0	0	4	1	4	1	0	0	4	0	4	0	4	4	4	3	1	3	1	4
C8	4	0	0	4	3	4	4	0	3	1	3	3	0	0	0	0	1	3	3	1
C9	1	1	1	1	4	3	0	1	0	4	3	0	4	3	3	3	3	0	3	4
C10	4	3	3	1	1	0	4	3	0	0	1	3	0	1	1	4	0	1	1	4
C11	4	0	4	0	3	0	0	1	1	3	0	4	0	0	4	0	4	1	4	1
C12	1	1	0	1	3	3	4	1	4	1	0	0	1	3	0	3	3	4	4	3
C13	0	3	3	3	4	0	0	4	0	4	4	3	0	1	1	1	4	1	0	1
C14	1	3	0	0	1	0	0	4	1	3	4	1	3	0	0	0	0	4	3	3
C15	0	4	1	3	1	3	0	4	1	3	0	4	3	4	0	1	3	0	4	3
C16	1	4	0	3	0	0	1	4	1	0	4	1	3	4	3	0	4	0	1	3
C17	0	1	0	0	3	3	3	3	1	4	0	1	0	4	1	0	0	1	1	0
C18	1	0	4	3	4	4	1	1	4	3	3	0	3	0	4	4	3	0	1	0
C19	4	4	3	4	0	1	3	1	1	3	0	0	4	1	0	3	3	3	0	1
C20	0	0	3	3	1	1	0	3	0	0	3	1	3	1	1	1	4	4	3	0

Note. CSF = critical success factor.

and

Table 4. Normalized influence matrix.

CSF	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17	C18	C19	C20
C1	0	0.004165	0.075737	0.058841	0.022920	0.022073	0.074780	0.005384	0.058206	0.004992	0.005254	0.057952	0.077158	0.060426	0.076366	0.058487	0.078408	0.057632	0.004286	0.076814
C2	0.075231	0	0.022560	0.058170	0.023317	0.003962	0.076023	0.075238	0.058750	0.022053	0.076723	0.057859	0.022133	0.022451	0.004588	0.004034	0.059815	0.076477	0.004164	0.076420
C3	0.003462	0.058382	0	0.076535	0.021652	0.057605	0.004445	0.076880	0.057834	0.022897	0.005179	0.075966	0.021978	0.077652	0.056439	0.075286	0.077116	0.003763	0.022483	0.022960
C4	0.020556	0.020484	0.003021	0	0.021917	0.002389	0.056735	0.003474	0.056816	0.057866	0.075511	0.056896	0.021008	0.075852	0.021158	0.020877	0.075759	0.021162	0.003645	0.022110
C5	0.057954	0.058932	0.057287	0.059398	0	0.074517	0.004262	0.022751	0.003284	0.058002	0.022418	0.022292	0.004692	0.059281	0.057038	0.076191	0.023546	0.003720	0.075784	0.059085
C6	0.058122	0.076426	0.022586	0.076722	0.003869	0	0.058246	0.005325	0.021702	0.076529	0.077920	0.022846	0.076839	0.077063	0.021925	0.075870	0.024473	0.004660	0.057593	0.059713
C7	0.002697	0.004573	0.076320	0.022936	0.076345	0.021778	0	0.005416	0.074880	0.005420	0.077014	0.004128	0.076668	0.077048	0.076662	0.058243	0.024244	0.056792	0.023061	0.076515
C8	0.075619	0.002741	0.004380	0.076212	0.057983	0.075388	0.076049	0	0.058124	0.022380	0.058145	0.056927	0.004610	0.004914	0.004287	0.005012	0.023019	0.057342	0.057903	0.022625
C9	0.021992	0.023101	0.022173	0.023012	0.075374	0.056923	0.003222	0.023315	0	0.077224	0.058623	0.004295	0.076072	0.058822	0.057402	0.058051	0.059357	0.003256	0.058606	0.076835
C10	0.074810	0.056646	0.057527	0.022622	0.021162	0.002803	0.075667	0.057815	0.004112	0	0.022083	0.057586	0.004019	0.022425	0.021627	0.075664	0.005243	0.022218	0.021247	0.076674
C11	0.074912	0.003619	0.075425	0.004236	0.057015	0.003974	0.003703	0.021677	0.021153	0.057719	0	0.075727	0.003232	0.004682	0.075144	0.004414	0.076685	0.020928	0.075787	0.021939
C12	0.021278	0.021831	0.004076	0.022684	0.058140	0.057650	0.075384	0.021705	0.075557	0.022773	0.005268	0	0.023362	0.058966	0.003967	0.058796	0.058852	0.075753	0.076031	0.058810
C13	0.004054	0.057154	0.057021	0.057882	0.075794	0.003536	0.004000	0.076269	0.003381	0.076111	0.075518	0.058593	0	0.022100	0.021015	0.021348	0.076601	0.021175	0.004010	0.021860
C14	0.021932	0.056234	0.003949	0.003962	0.021607	0.002539	0.003112	0.075350	0.020763	0.057450	0.075732	0.021285	0.056726	0	0.002659	0.002922	0.004348	0.075582	0.057181	0.057089
C15	0.004290	0.076441	0.020822	0.059038	0.021621	0.057159	0.005012	0.076969	0.021591	0.058886	0.005090	0.076069	0.057889	0.076561	0	0.021660	0.059074	0.004825	0.076212	0.058775
C16	0.021249	0.074875	0.002785	0.057725	0.003917	0.002571	0.021406	0.076679	0.021206	0.004363	0.076705	0.022324	0.057166	0.075801	0.056500	0	0.077101	0.004160	0.021842	0.057820
C17	0.003055	0.020897	0.002391	0.002853	0.056527	0.056434	0.056846	0.056606	0.020243	0.075086	0.003357	0.020126	0.002684	0.075020	0.019977	0.003179	0	0.020627	0.021187	0.003820
C18	0.021595	0.005229	0.075823	0.059258	0.075983	0.075923	0.021165	0.022764	0.075132	0.059708	0.058615	0.004687	0.058225	0.006334	0.076677	0.077059	0.059933	0	0.022519	0.005081
C19	0.074941	0.075430	0.057921	0.076750	0.003911	0.020521	0.058279	0.022463	0.022158	0.057914	0.005107	0.004125	0.076108	0.022969	0.003912	0.058081	0.059702	0.057551	0	0.022572
C20	0.002681	0.002883	0.056970	0.057961	0.021597	0.021237	0.002797	0.057435	0.003092	0.004371	0.057358	0.021327	0.056671	0.021451	0.020879	0.020973	0.076723	0.074522	0.057006	0

Note. CSF = critical success factor.

present the comprehensive results of the computational steps. First, using Formula (1), we constructed the direct relation matrix among the various factors (see Table 3). Subsequently, we normalized this matrix using Formula (2) to produce the standardized direct influence matrix (see Table 4). We then applied Formula (3) to calculate the total influence matrix (see

Table 5. Total influence matrix.

CSF	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17	C18	C19	C20
C1	0.097057	0.124720	0.195913	0.206523	0.153448	0.131925	0.189348	0.148721	0.176931	0.149766	0.151977	0.184871	0.214328	0.223540	0.197007	0.189299	0.255677	0.175638	0.132554	0.226335
C2	0.182057	0.099388	0.140868	0.195649	0.151485	0.108490	0.192580	0.202387	0.176852	0.155689	0.217645	0.178023	0.144864	0.165333	0.119214	0.125043	0.225614	0.194975	0.126202	0.217497
C3	0.106833	0.175177	0.096487	0.214500	0.137073	0.159935	0.119059	0.211928	0.168817	0.158534	0.143437	0.195995	0.141162	0.227349	0.157344	0.192674	0.236005	0.113639	0.144978	0.161728
C4	0.100557	0.104416	0.089071	0.094661	0.118216	0.077027	0.143656	0.102939	0.141963	0.166312	0.183092	0.148904	0.113039	0.190048	0.106352	0.112141	0.198928	0.108132	0.099268	0.131467
C5	0.167026	0.179252	0.167266	0.202178	0.104585	0.170929	0.117753	0.154274	0.107622	0.189739	0.157112	0.141559	0.127187	0.206163	0.161931	0.198763	0.182879	0.113551	0.196562	0.200917
C6	0.175536	0.202293	0.147087	0.225856	0.128541	0.094257	0.182446	0.148623	0.137427	0.222776	0.234431	0.154565	0.213092	0.233645	0.138499	0.205278	0.201194	0.127626	0.185359	0.214790
C7	0.104657	0.124732	0.196924	0.165319	0.202373	0.126458	0.097639	0.145150	0.183200	0.148155	0.221875	0.125829	0.206365	0.228110	0.195444	0.183902	0.192313	0.166150	0.153496	0.219728
C8	0.180195	0.104461	0.115407	0.209412	0.174180	0.175182	0.188078	0.107962	0.167190	0.151582	0.189164	0.166470	0.124612	0.143537	0.112119	0.126580	0.174754	0.162928	0.175816	0.155915
C9	0.131428	0.144269	0.136809	0.163184	0.193389	0.156255	0.112133	0.160444	0.097779	0.219821	0.197671	0.124554	0.200402	0.204242	0.165517	0.180566	0.221702	0.111081	0.185340	0.220171
C10	0.168557	0.153936	0.160485	0.147408	0.124737	0.089821	0.179934	0.175488	0.105002	0.105911	0.143525	0.165038	0.113369	0.150930	0.120497	0.186503	0.149091	0.126272	0.126598	0.205879
C11	0.167954	0.102496	0.175682	0.122948	0.157930	0.095651	0.102547	0.133387	0.114734	0.174644	0.099266	0.182533	0.105336	0.131229	0.171053	0.113089	0.217496	0.114791	0.185130	0.140886
C12	0.124424	0.135173	0.116801	0.159774	0.179302	0.160761	0.187169	0.147883	0.185565	0.159033	0.144725	0.105256	0.152324	0.204850	0.112511	0.185847	0.217166	0.189284	0.199852	0.199652
C13	0.102449	0.157764	0.156729	0.180377	0.186128	0.095086	0.107472	0.196733	0.099765	0.199269	0.195821	0.172294	0.093941	0.148398	0.115259	0.127342	0.219863	0.119123	0.112959	0.143194
C14	0.111539	0.141390	0.097360	0.112189	0.117394	0.079634	0.090599	0.179225	0.105414	0.164226	0.185504	0.114927	0.150569	0.097592	0.085540	0.095167	0.128743	0.169525	0.153784	0.163339
C15	0.114471	0.196236	0.128464	0.201851	0.139310	0.158734	0.125609	0.214396	0.131242	0.198790	0.144898	0.199037	0.181287	0.221931	0.096121	0.140396	0.219897	0.123375	0.202146	0.201149
C16	0.114514	0.172455	0.096174	0.176696	0.109377	0.086761	0.120199	0.197980	0.115892	0.122930	0.202501	0.130270	0.161562	0.200435	0.148529	0.091525	0.221395	0.105548	0.129393	0.178140
C17	0.076996	0.098344	0.076061	0.091885	0.136801	0.126269	0.136286	0.143743	0.091213	0.168439	0.095328	0.095452	0.082686	0.172639	0.089551	0.085629	0.093852	0.095946	0.104654	0.099645
C18	0.133391	0.132908	0.196379	0.209856	0.202958	0.187443	0.136047	0.163540	0.188824	0.211385	0.204718	0.133276	0.188487	0.165771	0.199019	0.212932	0.233267	0.102966	0.152818	0.152523
C19	0.176150	0.185531	0.169783	0.213431	0.118098	0.112420	0.172038	0.149721	0.130805	0.187326	0.139660	0.118833	0.197056	0.163317	0.109581	0.176928	0.219100	0.166217	0.103243	0.155845
C20	0.083031	0.088820	0.146109	0.168752	0.116374	0.103127	0.088265	0.160654	0.087315	0.113414	0.163568	0.114553	0.149200	0.131062	0.104558	0.113566	0.206775	0.161186	0.152378	0.096217

Note. CSF = critical success factor.

). Finally, utilizing Formulas (4)–(7), we computed each factor’s degree of influence, degree of being influenced, centrality, and rationality. These metrics are categorized and summarized in

Table 6. Factor analysis results: impact, centrality, and weighted rankings.

Factor	Influence	Affected	Centrality	Cause	Weight	Factor Attribute
C1	3.525579	2.618822	6.144401	0.906757	0.050090	Cause
C2	3.319856	2.823761	6.143617	0.496096	0.050084	Cause
C3	3.262655	2.805858	6.068513	0.456797	0.049472	Cause
C4	2.530190	3.462450	5.992640	−0.932260	0.048853	Effect
C5	3.247247	2.951697	6.198945	0.295550	0.050535	Cause
C6	3.573322	2.496166	6.069488	1.077155	0.049480	Cause
C7	3.387819	2.788855	6.176675	0.598964	0.050353	Cause
C8	3.105542	3.245180	6.350723	−0.139638	0.051772	Effect
C9	3.326758	2.713555	6.040313	0.613202	0.049242	Cause
C10	2.898982	3.367740	6.266721	−0.468758	0.051087	Effect
C11	2.808780	3.415919	6.224699	−0.607138	0.050745	Effect
C12	3.267351	2.952239	6.219590	0.315112	0.050703	Cause
C13	2.929967	3.060868	5.990835	−0.130901	0.048838	Effect
C14	2.543659	3.610120	6.153780	−1.066461	0.050167	Effect
C15	3.339342	2.705647	6.044989	0.633695	0.049280	Cause
C16	2.882278	3.043171	5.925448	−0.160893	0.048305	Effect
C17	2.161420	4.015713	6.177133	−1.854293	0.050357	Effect
C18	3.508506	2.747952	6.256458	0.760553	0.051004	Cause
C19	3.165085	3.022530	6.187615	0.142555	0.050443	Cause
C20	2.548924	3.485017	6.033942	−0.936093	0.049190	Effect

Note. CSF = critical success factor.

.

Navigating the causal structure reveals specific Critical Success Factors (CSFs) that drive the reduction in the digital divide. Specifically, the data identifies Scalable Technical Solutions (C1) and Cloud Computing Access (C2) as the primary ‘Driving CSFs’ (D − R > 0.3). Their high causality scores indicate that these technical infrastructures are the foundational triggers for the entire system, a finding consistent with empirical evidence that cloud computing enhances the efficiency of supply chain collaboration [44]. Conversely, Digital Investment Subsidies (C18) emerged as the ‘Core Intermediary CSF’ with the highest centrality score (D + R = 3.72), signaling that financial support is the most active hub for resource exchange. Furthermore, while the Collaborative Work Environment Platform (C20) is identified as a ‘Result CSF’, its strong connection to other factors suggests it acts as the essential testing ground where digital governance effectiveness is ultimately realized, which aligns with evidence supporting its role in enhancing system responsiveness [63].

Further analysis reveals that C1 and C2 exert a significant positive driving effect on technical and managerial factors such as mobile technology integration (C3), interoperable systems (C5), and resource integration platforms (C14), thereby validating the three-stage theoretical framework of “technology-driven—organizational synergy—institutional safeguards” [64]. Empowerment-related factors, including digital transformation leadership training (C10) and skill upgrade incentives (C12), primarily function as the “capability-bearing end” of the system, with their effectiveness contingent upon the driving force generated by front-end technology deployment.

Across all configurations, C1 consistently emerges as the most influential driving factor in the system $(D-R=0.73)$, while C18 remains the most central and densely connected node. In contrast, factors such as C15 and C8 are often in a passive response state $(D-R<-0.4)$. The systems’ overall average centrality is $\stackrel{-}{\mathrm{E}}=$ $3.27$, with a standard deviation of $\sigma =0.36$, reflecting a well-coupled three-stage configuration of “technology leadership—organizational synergy—institutional safeguards.” This structure suggests a high degree of stability and coordination in the digital divide governance framework [63].

To enhance the visual representation of the causal structure, we present a bar chart based on each key success factor’s causality degree ${(D-R)}_i$ (see Figure 2). A positive ${(D-R)}_i$ value indicates that a factor has a strong driving effect (first and second quadrants), whereas a negative ${(D-R)}_i$ value suggests that the factor functions primarily as a response variable (third and fourth quadrants). The bar chart provides an intuitive view of each factor’s positioning and behavioral tendencies within the system. For example, factors F6 and F1 exhibit significant driving characteristics, whereas factors F17 and F14, with negative causality degrees, display a strong reactive nature. This illustration complements the two-dimensional causal relationship diagram and supports further importance analysis and strategic decision-making.

In Figure 3, each point represents a key success factor, with its causal position $(D-RD-RD-R)$ and centrality $(D+RD+RD+R)$ determining its role within the causal network. Factors in the upper-right quadrant are core drivers and should receive priority in governance strategies, whereas those in the lower-left quadrant function as response-oriented factors with relatively limited influence.

4.2. Analytic Network Process Results

Table 7. Normalization matrix.

CSF	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17	C18	C19	C20
C1	0	0.000000	0.114286	0.068182	0.027778	0.032258	0.117647	0.000000	0.088235	0.000000	0.000000	0.083333	0.102564	0.066667	0.117647	0.076923	0.080000	0.090909	0.000000	0.090909
C2	0.125000	0	0.028571	0.068182	0.027778	0.000000	0.117647	0.102564	0.088235	0.024390	0.095238	0.083333	0.025641	0.022222	0.000000	0.000000	0.060000	0.121212	0.000000	0.090909
C3	0.000000	0.085714	0	0.090909	0.027778	0.096774	0.000000	0.102564	0.088235	0.024390	0.000000	0.111111	0.025641	0.088889	0.088235	0.102564	0.080000	0.000000	0.027027	0.022727
C4	0.031250	0.028571	0.000000	0	0.027778	0.000000	0.088235	0.000000	0.088235	0.073171	0.095238	0.083333	0.025641	0.088889	0.029412	0.025641	0.080000	0.030303	0.000000	0.022727
C5	0.093750	0.085714	0.085714	0.068182	0	0.129032	0.000000	0.025641	0.000000	0.073171	0.023810	0.027778	0.000000	0.066667	0.088235	0.102564	0.020000	0.000000	0.108108	0.068182
C6	0.093750	0.114286	0.028571	0.090909	0.000000	0	0.088235	0.000000	0.029412	0.097561	0.095238	0.027778	0.102564	0.088889	0.029412	0.102564	0.020000	0.000000	0.081081	0.068182
C7	0.000000	0.000000	0.114286	0.022727	0.111111	0.032258	0	0.000000	0.117647	0.000000	0.095238	0.000000	0.102564	0.088889	0.117647	0.076923	0.020000	0.090909	0.027027	0.090909
C8	0.125000	0.000000	0.000000	0.090909	0.083333	0.129032	0.117647	0	0.088235	0.024390	0.071429	0.083333	0.000000	0.000000	0.000000	0.000000	0.020000	0.090909	0.081081	0.022727
C9	0.031250	0.028571	0.028571	0.022727	0.111111	0.096774	0.000000	0.025641	0	0.097561	0.071429	0.000000	0.102564	0.066667	0.088235	0.076923	0.060000	0.000000	0.081081	0.090909
C10	0.125000	0.085714	0.085714	0.022727	0.027778	0.000000	0.117647	0.076923	0.000000	0	0.023810	0.083333	0.000000	0.022222	0.029412	0.102564	0.000000	0.030303	0.027027	0.090909
C11	0.125000	0.000000	0.114286	0.000000	0.083333	0.000000	0.000000	0.025641	0.029412	0.073171	0	0.111111	0.000000	0.000000	0.117647	0.000000	0.080000	0.030303	0.108108	0.022727
C12	0.031250	0.028571	0.000000	0.022727	0.083333	0.096774	0.117647	0.025641	0.117647	0.024390	0.000000	0	0.025641	0.066667	0.000000	0.076923	0.060000	0.121212	0.108108	0.068182
C13	0.000000	0.085714	0.085714	0.068182	0.111111	0.000000	0.000000	0.102564	0.000000	0.097561	0.095238	0.083333	0	0.022222	0.029412	0.025641	0.080000	0.030303	0.000000	0.022727
C14	0.031250	0.085714	0.000000	0.000000	0.027778	0.000000	0.000000	0.102564	0.029412	0.073171	0.095238	0.027778	0.076923	0	0.000000	0.000000	0.000000	0.121212	0.081081	0.068182
C15	0.000000	0.114286	0.028571	0.068182	0.027778	0.096774	0.000000	0.102564	0.029412	0.073171	0.000000	0.111111	0.076923	0.088889	0	0.025641	0.060000	0.000000	0.108108	0.068182
C16	0.031250	0.114286	0.000000	0.068182	0.000000	0.000000	0.029412	0.102564	0.029412	0.000000	0.095238	0.027778	0.076923	0.088889	0.088235	0	0.080000	0.000000	0.027027	0.068182
C17	0.000000	0.028571	0.000000	0.000000	0.083333	0.096774	0.088235	0.076923	0.029412	0.097561	0.000000	0.027778	0.000000	0.088889	0.029412	0.000000	0	0.030303	0.027027	0.000000
C18	0.031250	0.000000	0.114286	0.068182	0.111111	0.129032	0.029412	0.025641	0.117647	0.073171	0.071429	0.000000	0.076923	0.000000	0.117647	0.102564	0.060000	0	0.027027	0.000000
C19	0.125000	0.114286	0.085714	0.090909	0.000000	0.032258	0.088235	0.025641	0.029412	0.073171	0.000000	0.000000	0.102564	0.022222	0.000000	0.076923	0.060000	0.090909	0	0.022727
C20	0.000000	0.000000	0.085714	0.068182	0.027778	0.032258	0.000000	0.076923	0.000000	0.000000	0.071429	0.027778	0.076923	0.022222	0.029412	0.025641	0.080000	0.121212	0.081081	0

Note. CSF = critical success factor.

presents the normalized matrix, which integrates the aggregated clarity values reflecting the interdependent effects among dimensions and indicators. Based on this integration, we utilized Formula (8) to generate the converged limit matrix and subsequently ranked the weights of these dimensions and indicators.

From the perspective of the overall research question, the ANP weights provide a quantitative hierarchy of the CSFs. As detailed in the limit supermatrix (see

Table 8. Convergent limited supermatrix.

CSF	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17	C18	C19	C20
C1	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818	0.056818
C2	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743	0.054743
C3	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186	0.052186
C4	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212	0.040212
C5	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368	0.053368
C6	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341	0.056341
C7	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689	0.054689
C8	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830	0.053830
C9	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549	0.052549
C10	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882	0.048882
C11	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591	0.047591
C12	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437	0.054437
C13	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469	0.046469
C14	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250	0.042250
C15	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229	0.053229
C16	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663	0.044663
C17	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741	0.036741
C18	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860	0.057860
C19	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169	0.052169
C20	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974	0.040974

Note. CSF = critical success factor.

), Scalable Technical Solutions (C1) and Cloud Computing Access (C2) secured the highest global weights of 5.68% and 5.47%, respectively. This statistical dominance confirms their status as the ‘Foundational CSFs’ required to bridge the infrastructure gap. The analysis further highlights a tiered structure where technology deployment factors outrank soft governance measures. For instance, Digital Transformation Leadership Training (C10), despite its theoretical importance, ranked 13th (4.76%). This data disparity suggests that in the current stage of the GCMRS, closing the ‘hardware divide’ is a prerequisite for the effectiveness of ‘soft power’ CSFs like leadership and localization.

Regardless of how the ANP network structure evolves, the four core variables—C1 (scalable solutions), C2 (cloud computing), C18 (digital investment subsidies), and C20 (collaborative platforms)—maintain a highly consistent weighting order with minimal fluctuations. This stability highlights a structural feature of GCMRS digital governance in which technological support dominates the decision-making framework, thereby compressing the weights of certain supporting institutional and collaborative factors, such as C10 (leadership training) and C15 (localized content), to below 5%.

As decision-making networks have become more complex, the average weight of front-end technical factors has increased by 11.2%, while management and empowerment-related factors marginally declined, particularly among non-core institutional variables. Although C10 demonstrates high centrality in the fuzzy DEMATEL analysis (D + R = 3.55), it ranks only 13th in the ANP results, further illustrating a “deployment-first, governance-lagging” dynamic.

The average weight of all key factors is W⁻ = 5.26%, and the coefficient of variation is CV = 7.9%, which is significantly lower than the robustness threshold (CV < 15%), confirming the high consistency and stability of the constructed multi-level decision-making network.

Figure 4 presents the CSF network structure diagram based on the causal pathways identified in the fuzzy DEMATEL analysis. The arrows indicate the direction of influence, illustrating the direct dependencies among factors. This diagram serves as the structural foundation for the ANP supermatrix and final weight extraction and reveals the complex interaction patterns that characterize multi-factor coupling and evolution in the digital governance of critical mineral resource supply chains.

4.3. Choquet Integral Analysis

To identify the CSFs that reduce the digital divide among partners in the GCMRS, we applied the Choquet integral method (Formulas (9)–(10)) to capture the potential non-additive preference relationships among factors and to establish a more realistic preference-based ranking system. Drawing upon the λ − additivity principle in Formula (9) and the fuzzy integral formulation in Formula (10), we used the stabilized weights from the finalized ANP supermatrix (see Table 5), together with the fuzzy measure values (see

Table 9. Single cluster with final weighting.

CSF	Final Weight
C1	0.056818
C2	0.054743
C3	0.052186
C4	0.040212
C5	0.053368
C6	0.056341
C7	0.054689
C8	0.053830
C9	0.052549
C10	0.048882
C11	0.047591
C12	0.054437
C13	0.046469
C14	0.042250
C15	0.053229
C16	0.044663
C17	0.036741
C18	0.057860
C19	0.052169
C20	0.040974

Note. CSF = critical success facto.

), to conduct Choquet integral-based synthesis and reassess each factor’s relative significance. This process produced a unified prioritization cluster. The final Choquet-adjusted weights (see Table 8) cover all 20 CSFs and realistically reflect their synergistic significance, thereby supporting preference-enhanced decision-making.

With $\lambda$ ranging between −0.5 and 0.5, the Choquet integral value of the technology deployment-oriented combination (C1 + C2 + C6) increases from 36.814 to 58.209—a 58.1% increase—while the capability–collaboration synergy combination (C18 + C20 + C10) fluctuates by only 1.9% (from 46.217 to 47.103), indicating strong stability and robust synergy. The most effective optimization path occurs at λ = −0.25, where the synergy combination achieves a 24.3% gain. C18 (digital investment subsidies), C20 (collaborative platforms), and C10 (digital transformation leadership training) form its core and removing C20 significantly weakens the path’s effectiveness, underscoring the critical role of collaborative platforms in digital governance. In contrast, localized content (C15) and supply chain environmental data monitoring tools (C8) are absent from any high-efficiency combinations and contribute less than 4.1% at most, confirming their marginal strategic role. The detailed Choquet integral calculation results are shown (see

Table 10. Choquet integral results.

λ Value	Traceability	Visibility	Disclosure	Openness
−0.50	0.828	0.828	0.828	0.828
−0.49	0.831	0.831	0.831	0.831
−0.48	0.834	0.834	0.834	0.834
−0.47	0.838	0.838	0.838	0.838
−0.46	0.841	0.841	0.841	0.841
−0.45	0.844	0.844	0.844	0.844
−0.44	0.847	0.847	0.847	0.847
−0.43	0.850	0.850	0.850	0.850
−0.42	0.854	0.854	0.854	0.854
−0.41	0.857	0.857	0.857	0.857
−0.40	0.860	0.860	0.860	0.860
−0.39	0.863	0.863	0.863	0.863
−0.38	0.867	0.867	0.867	0.867
−0.37	0.870	0.870	0.870	0.870
−0.36	0.873	0.873	0.873	0.873
−0.35	0.877	0.877	0.877	0.877
−0.34	0.880	0.880	0.880	0.880
−0.33	0.883	0.883	0.883	0.883
−0.32	0.887	0.887	0.887	0.887
−0.31	0.890	0.890	0.890	0.890
−0.30	0.893	0.893	0.893	0.893
−0.29	0.897	0.897	0.897	0.897
−0.28	0.900	0.900	0.900	0.900
−0.27	0.904	0.904	0.904	0.904
−0.26	0.907	0.907	0.907	0.907
−0.25	0.910	0.910	0.910	0.910
−0.24	0.914	0.914	0.914	0.914
−0.23	0.917	0.917	0.917	0.917
−0.22	0.921	0.921	0.921	0.921
−0.21	0.924	0.924	0.924	0.924
−0.20	0.928	0.928	0.928	0.928
−0.19	0.931	0.931	0.931	0.931
−0.18	0.935	0.935	0.935	0.935
−0.17	0.938	0.938	0.938	0.938
−0.16	0.942	0.942	0.942	0.942
−0.15	0.945	0.945	0.945	0.945
−0.14	0.949	0.949	0.949	0.949
−0.13	0.952	0.952	0.952	0.952
−0.12	0.956	0.956	0.956	0.956
−0.11	0.960	0.960	0.960	0.960
−0.10	0.963	0.963	0.963	0.963
−0.09	0.967	0.967	0.967	0.967
−0.08	0.970	0.970	0.970	0.970
−0.07	0.974	0.974	0.974	0.974
−0.06	0.978	0.978	0.978	0.978
−0.05	0.981	0.981	0.981	0.981
−0.04	0.985	0.985	0.985	0.985
−0.03	0.989	0.989	0.989	0.989
−0.02	0.993	0.993	0.993	0.993
−0.01	0.996	0.996	0.996	0.996
0.00	1.500	1.500	1.500	1.500
0.01	1.004	1.004	1.004	1.004
0.02	1.008	1.008	1.008	1.008
0.03	1.011	1.011	1.011	1.011
0.04	1.015	1.015	1.015	1.015
0.05	1.019	1.019	1.019	1.019
0.06	1.023	1.023	1.023	1.023
0.07	1.026	1.026	1.026	1.026
0.08	1.030	1.030	1.030	1.030
0.09	1.034	1.034	1.034	1.034
0.10	1.038	1.038	1.038	1.038
0.11	1.042	1.042	1.042	1.042
0.12	1.046	1.046	1.046	1.046
0.13	1.050	1.050	1.050	1.050
0.14	1.054	1.054	1.054	1.054
0.15	1.057	1.057	1.057	1.057
0.16	1.061	1.061	1.061	1.061
0.17	1.065	1.065	1.065	1.065
0.18	1.069	1.069	1.069	1.069
0.19	1.073	1.073	1.073	1.073
0.20	1.077	1.077	1.077	1.077
0.21	1.081	1.081	1.081	1.081
0.22	1.085	1.085	1.085	1.085
0.23	1.089	1.089	1.089	1.089
0.24	1.093	1.093	1.093	1.093
0.25	1.097	1.097	1.097	1.097
0.26	1.101	1.101	1.101	1.101
0.27	1.106	1.106	1.106	1.106
0.28	1.110	1.110	1.110	1.110
0.29	1.114	1.114	1.114	1.114
0.30	1.118	1.118	1.118	1.118
0.31	1.122	1.122	1.122	1.122
0.32	1.126	1.126	1.126	1.126
0.33	1.130	1.130	1.130	1.130
0.34	1.134	1.134	1.134	1.134
0.35	1.139	1.139	1.139	1.139
0.36	1.143	1.143	1.143	1.143
0.37	1.147	1.147	1.147	1.147
0.38	1.151	1.151	1.151	1.151
0.39	1.155	1.155	1.155	1.155
0.40	1.160	1.160	1.160	1.160
0.41	1.164	1.164	1.164	1.164
0.42	1.168	1.168	1.168	1.168
0.43	1.172	1.172	1.172	1.172
0.44	1.177	1.177	1.177	1.177
0.45	1.181	1.181	1.181	1.181
0.46	1.185	1.185	1.185	1.185
0.47	1.190	1.190	1.190	1.190
0.48	1.194	1.194	1.194	1.194
0.49	1.198	1.198	1.198	1.198
0.50	1.203	1.203	1.203	1.203
∫ hdg (λ < 0)	45.493	45.493	45.493	45.493
∫ hdg (λ > 0)	56.535	56.535	56.535	56.535
∫ hdg (Overall)	102.028	102.028	102.028	102.028
Avg ∫ hdg (λ < 0)	0.910	0.910	0.910	0.910
Avg ∫ hdg (λ > 0)	1.109	1.109	1.109	1.109
Avg ∫ hdg (Overall)	1.010	1.010	1.010	1.010

).

No matter how λ changes, the performance ranking of the four key combinations remains stable: technology deployment oriented (C1 + C2 + C6), capability and collaboration synergy (C18 + C20 + C10), multi-level standards promotion (C5 + C17 + C14), and governance support enhancement (C10 + C14 + C16). Among these, C18 + C20 + C10 consistently delivers the best performance. As λ increases, the effectiveness of technology-related factors, such as C1 and C2, grows, while for λ < 0, governance and collaboration paths become more dominant. For example, C10’s relative contribution rises significantly when λ = −0.5.

The average integral value across all combinations is $\mathrm{I}$ = 48.93, and the coefficient of variation is CV = 4.9%, well below the robustness threshold. The proposed three-tier model of deployment, capability, and collaboration demonstrates strong adaptability across regions and stakeholders, offering robust support for digital collaborative governance in the GCMRS.

5. Discussion

5.1. Theoretical Contribution

This study contributes to theory in three key aspects. First, it enriches the theoretical perspective on the governance of critical mineral resource production supply chains 4.0. Second, it expands the research boundaries of digital supply chain management. Third, it deepens the theoretical understanding of collaborative mechanisms among global supply chain partners. Specifically, this study adopts a three-stage approach that integrates fuzzy DEMATEL, ANP, and Choquet integral to identify CSFs with strong causal strengths and high importance. By systematically comparing its results with those of existing studies, the paper highlights its theoretical innovation and academic value.

First, this paper introduces a novel research perspective into critical mineral resource production supply chain management 4.0. While existing studies primarily examine the spatial optimization of resource allocation [65] and the effects of trade barriers on supply chain stability [66], few address disparities in digital capabilities among supply chain partners and their impact on collaborative efficiency [67]. This study identifies key drivers such as “technical infrastructure development” (C6), “digital governance capabilities” (C15), and “information system interoperability” (C10) as central nodes within the causal structure. The findings validate the argument that “insufficient digital interoperability serves as a bottleneck for collaboration” [68]. Moreover, the study extends the conceptual framework of corporate digital resilience to encompass collaborative governance mechanisms among heterogeneous actors within global critical mineral resource supply networks 4.0 [17].

Second, this paper broadens the research boundaries of digital supply chain management. While existing literature predominantly focuses on digital transformation in manufacturing and retail sectors [69,70], it pays limited attention to cross-border digital collaboration in high-uncertainty industries such as mining. In contrast, this study integrates ANP and Choquet integral analysis to identify “data standardization level” (C2), “establishment of key data sharing platforms” (C7), and “security assurance mechanisms” (C18) as core elements ranking highest in weight-based evaluations. These results underscore their pivotal roles in digital supply chain governance. Unlike Horváth and Szabó’s [69] assertion that “technology deployment precedes organizational coordination” (p. 8), this study highlights that inter-organizational data sharing and digital infrastructure security serve as fundamental prerequisites for improving collaboration efficiency in critical mineral supply chains. These findings extend the applicability of digital governance models to resource-oriented supply chains and address the research gap on digital collaboration mechanisms between resource-exporting and resource-consuming countries within global value chains.

Third, this paper enriches the theoretical perspective on global supply chain partner management. Existing studies primarily analyze collaborative mechanisms from the viewpoints of governance structures [71], organizational trust [72], and knowledge transfer [73], focusing on power dynamics and institutional arrangements between lead firms and subordinate partners [74]. However, researchers have yet to conduct quantitative investigations into partner heterogeneity in digital capabilities and resource allocation, along with corresponding strategic responses. This study incorporates the Choquet integral method to model fuzzy preference relationships among factors such as “digital capability training programs” (C8), “leadership digital cognition level” (C14), and “cross-border coordination mechanisms” (C11). The model reveals the interdependent structures and complexity of collaboration among supply chain actors. By capturing the dynamic interactions of digital collaboration behaviors among heterogeneous multinational organizations from a quantitative perspective, this study provides novel modeling insights and explanatory frameworks for advancing the theory of global supply chain 4.0 partnership management.

Finally, this paper extends the applicability of Innovation Diffusion Theory, the Resource-Based View, and Social Network Theory to the governance of the digital divide. By systematically analyzing the causal weights of key factors such as Technical Infrastructure Construction (C6), Cloud Computing Access (C2), and Digital Governance Capacity (C15), the study confirms the theoretical postulates regarding technological accessibility and resource endowment. Specifically, the high centrality of subsidy-related factors validates the RBV perspective on resource orchestration. Simultaneously, the prominence of collaborative platforms supports the connectivity arguments of Social Network Theory. This integration highlights both the validity and the practical boundaries of these frameworks when applied to complex global supply chain contexts.

5.2. Practical Implication

Drawing on the CSFs identified in this study, we propose targeted managerial recommendations for three stakeholder groups: governments of resource-rich and resource-consuming countries, critical mineral resource supply chain enterprises, and industry practitioners. These recommendations move beyond generic digitalization goals to provide a quantifiable basis for decision-making, specifically addressing the technical and operational asymmetries that currently fragment the GCMRS.

At the governmental level, stakeholders should prioritize infrastructure development and institutional guidance to enhance system connectivity and visibility. Governments of resource-rich and resource-consuming countries should strengthen the coordinated construction of cross-border digital infrastructure to facilitate interconnected data transmission channels and platform deployment (C6). They should establish unified interface protocols and data exchange standards to improve platform interoperability and system integration efficiency (C10) [75]. In addition, they should deploy policy instruments such as fiscal subsidies and tax incentives to encourage enterprises to engage in the joint construction and sharing of data platforms and information disclosure mechanisms, thereby enhancing cross-border collaborative regulatory capacity and overall system transparency.

At the enterprise level, supply chain actors should focus on capacity building and enhancing collaborative mechanisms that serve as core drivers of digital cooperation. Enterprises should build systematic digital governance capabilities (C15) and strengthen organizational adaptability in terms of system compatibility, standards interoperability, and cross-border operations through targeted employee training programs (C8). In cross-border collaborations, they should establish stable information-sharing mechanisms (C7), proactively engage in regional cooperation networks, and participate in industry-level interface standards development to reduce coordination costs across heterogeneous platforms. They should also prioritize cybersecurity safeguards (C18), integrating institutional and technical measures to improve data security and system resilience, thereby minimizing potential digital risks.

At the industry practitioner level, actors should bridge diverse stakeholders and provide tailored support to enterprises exhibiting significant capability gaps. Technical service institutions should advance the modular development of core platforms and tools, offering flexible middleware solutions that lower technical entry barriers for enterprise-level digital system integration (C10). They should also establish a standardized digital capability rating and certification framework to facilitate resource allocation and service alignment, thereby enhancing the overall maturity of digital collaboration across the industry. Practitioners should reinforce knowledge dissemination mechanisms (C13) through cross-regional case studies, targeted seminars, and joint innovation projects to foster experience sharing and capability convergence between resource-rich and consumer countries. These efforts will ultimately improve the adaptability and resilience of the global critical mineral resource supply chain.

In summary, all stakeholders should, according to their respective roles, prioritize system interconnectivity, capacity enhancement, information sharing, and platform standardization. Through coordinated efforts, they can build an inclusive, interoperable, and efficient global digital governance framework for the production of critical mineral resources. Such a framework will play a pivotal role in narrowing digital capability disparities and enhancing cross-border collaboration efficiency in the global supply chain.

6. Conclusions

This study tackles the digital divide among partners in the Global Critical Mineral Supply Chain (GCMRS) by constructing a three-stage analytical framework that integrates fuzzy DEMATEL, the analytic network process (ANP), and Choquet integral methods. Leveraging expert questionnaire data, we systematically identified and evaluated the critical success factors (CSFs) that influence the resolution of digital disparities across supply chain partners. The analysis examines three dimensions—causal structure, weight ranking, and preference coupling—to reveal the underlying logic of key factors.

The findings indicate that Blockchain technology for supply chain transparency (C6), Cloud Computing Access (C2), Digital Investment Subsidy (C18), and a Collaborative Work Environment Platform (C20) are the most influential core variables, distinguished by their strong causal effects and high comprehensive weights. Specifically, the fuzzy DEMATEL analysis highlights the pronounced causal driving effects of Blockchain technology for supply chain transparency (C6) and Cloud Computing Access (C2), while ANP results show that Digital Investment Subsidy (C18) and a Collaborative Work Environment Platform (C20) hold significantly higher weights compared to other variables. Their associated tertiary indicators, including digital literacy training, multi-level digital standardization, and technical support for capacity building, also received notably higher scores. Moreover, the Choquet integral analysis reveals preference-enhancing interactions among these factors, indicating their synergistic influence in practical digital governance scenarios.

Overall, this study addresses the lack of quantitative prioritization in current digital governance models. Its findings are expected to provide a quantifiable basis for decision-making in policy formulation or industry practices. Specifically, it enables managers to optimize platform mechanisms and resource allocation based on the identified ‘technology-first’ causal hierarchy.

Despite its contributions, the study has certain limitations. First, it focuses specifically on the key mineral resource industry, and the unique characteristics and cooperative structures of this sector may constrain the generalizability of findings to other industries. Second, although the expert scoring approach ensures a degree of professional judgment, the assessment of causal relationships and factor weights may still reflect individual knowledge backgrounds and subjective biases. Third, this study primarily addresses the digital representation layer (e.g., digital twins, blockchain) of the supply chain, with less emphasis on the physical characteristics of the critical materials themselves. As highlighted by recent research on the ‘Internet of Materials’ [76], bridging the gap between digital traceability and physical material properties is essential for achieving true circularity. Future research should therefore aim to incorporate these physical attributes into the proposed model to enhance its robustness. Lastly, the theoretical integration in this study centers primarily on three mainstream perspectives—innovation diffusion, resource-based view, and social network theory—without incorporating complementary frameworks such as institutional evolution or cultural embeddedness. Future studies could broaden the theoretical scope to explore the dynamic mechanisms underlying collaborative factors in cross-institutional and culturally diverse contexts.