Research on the Green and Low-Carbon Development Path of Digital Intelligence Empowering Enterprises in Manufacturing Industry

Abstract

The Fourth Industrial Revolution, driven by advancements in information technology, has ushered humanity into the age of intelligence. As digital technologies like artificial intelligence and large-scale models continue to evolve and gain traction, the convergence of digital innovation and green development within manufacturing enterprises has emerged as a pivotal trend. This integration not only fosters high-quality, sustainable growth, but also increasingly validates the impact of digital intelligence on advancing low-carbon performance. This study delves into how manufacturing enterprises can attain sustainable and low-carbon growth via digital transformation, employing the entropy TOPSIS evaluation model to assess the effectiveness of various empowerment strategies. Based on the findings, the paper offers actionable recommendations for enhancing sustainable practices in manufacturing during this digital shift. Beyond enriching the theoretical framework on the synergy between digital intelligence and sustainability in manufacturing, this research provides practical insights and guidance for enterprises leveraging next-generation digital technologies to drive their green and low-carbon initiatives more effectively.

1. Introduction

Integrating digital intelligence and greening has become a significant trend in advancing high-quality and sustainable development for enterprises [1]. China’s 20th National Congress of the Communist Party made it crystal clear: they are all in on digital advancement, a swift transition to eco-friendly practices, and the push for manufacturing that is cutting-edge, smart, and green. Echoing this sentiment, the 14th Five-Year Plan for Green Industrial Development hammered home the idea that green, low-carbon growth needs to be treated as a comprehensive undertaking and fast-tracked accordingly. This demonstrates that promoting the integration of digital intelligence and greening in enterprises has become a national strategy. From an international perspective, the American government has issued the “American Manufacturing Innovation Network” plan, aiming to improve the competitiveness and green level of the manufacturing industry through digital intelligence technology [2]. The European Union rolled out its “European Green New Deal” and “Digital Strategy”, aiming to steer businesses toward sustainable industrial transformation by merging digital intelligence with eco-friendly innovations. Meanwhile, Japan has crafted its “Green Growth Strategy”, outlining a road map to drive industries toward low-carbon and environmentally conscious practices by effortlessly combining digital innovations with eco-friendly solutions. Both initiatives underscore the global shift toward harmonizing technological advancement with ecological responsibility [3]. As the economic foundation in China, the manufacturing sector is crucial for leveraging advanced digital technologies to modernize traditional industries. Harnessing the seamless fusion of digital intelligence and sustainable practices within the manufacturing sector is a cornerstone for meeting China’s “dual carbon” objectives. This integration accelerates the nation’s shift from traditional industrial paradigms to a more ecologically conscious civilization and paves the way for a greener, more innovative future [4]. Driving the convergence of digital intelligence and sustainability within the manufacturing sector is far more than a straightforward evolution—it is a paradigm shift. Cutting-edge digital innovations like artificial intelligence and big data must be seamlessly woven into every stage of the traditional manufacturing lifecycle, from production to management. This is not just about incremental change; it is about redefining the entire process for a smarter, greener future [5], realizing the integration of digital intelligence and greening, and realizing the deep integration of “number” and “green” by reducing carbon emissions (as illustrated in Figure 1) [6].

The existing research literature confirms that enterprise digital transformation effectively promotes low-carbon, green, and sustainable development—a consensus widely recognized among business managers, practitioners, and scholars [7]. For instance, Li et al. (2024) examined the relationship between digital transformation and green transformation in manufacturing companies, concluding that digital technologies serve as key drivers of sustainable industrial upgrading and ecological integration [8]. Similarly, Lu et al. (2023) explored the internal mechanisms through which digital transformation facilitates green transitions, identifying a structured pathway that progresses from basic digital adoption to intelligent and ecologically integrated manufacturing systems. Their findings suggest that digital transformation enables enterprises to achieve green structuralization, capability-building, and leverage for long-term sustainability [9]. Moreover, Wang et al. (2023) analyzed the mechanism of structural optimization in digital transformation and its effects on green total factor productivity, providing empirical evidence that digital advancements support green manufacturing development [10].

Foss et al. (2017) noted that digital transformation allows enterprises to break free from a single-dimensional growth model [11]. By leveraging digital technologies, enterprises can secure more excellent development space in dynamic markets, stimulate advanced green innovation, and accomplish enhanced green transition outcomes [11]. Gallo et al. (2023) argued that the integration of big data analytics (BDA), artificial intelligence (AI), and cloud computing significantly improves corporate environmental performance and fosters green supply chain innovation [12]. Their findings highlighted that data-driven technologies enable firms to optimize resource utilization, improve energy efficiency, and drive sustainability in operational processes [12]. Regarding investigating the efficiency of digital intelligence transformation in manufacturing enterprises and its implementation pathways, Zhao and Li (2024) explored the relationship between digital transformation and corporate innovation performance, revealing that this relationship follows an inverted U-shaped pattern [13]. They found that while digitalization enhances innovation performance up to a certain point, excessive digitalization can lead to diminishing returns. Furthermore, their study identified human capital as a key mediating factor, reinforcing the role of knowledge and skills in optimizing innovation outcomes [13]. Paavola et al. (2017) argued that enterprise digitalization represents a fundamental transformation encompassing business activities, operational processes, organizational capabilities, and business models [14]. Its main goal and effectiveness are centered on improving customer experience, optimizing operations, and supporting the creation of innovative business models [14]. Xue et al. (2024) examined the digital transformation strategy of manufacturing enterprises, emphasizing its role in process-driven value enhancement and customer-oriented value creation [15]. Their study highlighted how companies strategically navigate digital transformation by aligning with local and global digitalization frameworks while optimizing their competitive advantages through either product-centric or service-oriented approaches [15]. Mubarak et al. (2020) examined influences on digital transformation within the manufacturing industry, concentrating on technology, organizational structure, and strategic direction [16]. Thiede (2021) explored the key drivers and transformation pathways of digitalization in the manufacturing sector, centering on industrial enterprises’ digital technologies, methodologies, and tools [17]. Duraivelu (2022) argued that technological innovation and organizational culture are critical factors influencing enterprise digital transformation, as evidenced by an analysis of the interplay between technology and organizational dynamics [18].

Within the study of the convergence between digital intelligence transformation and corporate green development, Warner et al. (2019) analyzed how digital technologies help enterprises overcome resource barriers, effectively integrate internal and external resources, optimize resource allocation, and improve green economic efficiency [19]. Carrión-Flores et al. (2010) thought that an enterprise’s digital intelligence promotes an enterprise’s green innovation, which is of great significance to an enterprise’s environmental performance and sustainable development [20]. Li and Hou (2024) analyzed evolutionary trends in the integration and coordinated development of digitalization and green transformation within the manufacturing sector [21]. Their study highlighted that digitalization enhances resource efficiency, optimizes industrial structures, and accelerates green upgrading processes, ultimately promoting the sustainable transformation of manufacturing enterprises [21]. Additionally, Khan et al. (2022) analyzed how digital transformation fosters sustainable supply chain performance through smart technologies, emphasizing the role of AI, big data, and cloud computing in driving green innovation. Their study demonstrated that digital transformation helps balance green transition and industrial upgrading by improving data efficiency and reducing carbon footprints [22]. Yin and Yu (2022) explored how digital green knowledge adoption enhances green innovation practices in Industry 5.0 [23]. Their study highlighted the role of AI, big data, and cloud computing in optimizing energy consumption and promoting intelligent decision-making for sustainable industrial growth [23]. Liu Jun et al. (2025) emphasized that manufacturing enterprises’ digital-intelligent transformation is essential to enhancing low-carbon performance [24]. This improvement is achieved by reinforcing enterprise technological innovation, advancing traditional industry transformation and upgrading, and optimizing enterprise resource allocation [24]. Di et al. (2024) analyzed how digital empowerment drives green and low-carbon industrial development, highlighting regional variations in its effectiveness [25]. Their study identified key mechanisms through which digital transformation enhances energy efficiency and reduces carbon emissions in industrial enterprises, emphasizing the role of intelligent digital platforms [25]. Their research conclusion showed that digital empowerment for greening has evolved in stages, and in the process of digital empowerment for greening and low carbon, three-stage digital capabilities can effectively drive enterprises to complete the evolution of a green structure, green orientation, and green demonstrations.

Based on the findings of the aforementioned scholars, substantial and well-established research has been conducted on the efficiency and implementation pathways of enterprise digital intelligence transformation, the enhancement of green performance and sustainable development, and the integration of digital intelligence transformation with developing in a green and low-carbon manner. However, existing studies have rarely addressed the efficiency measurement of how adopting intelligent digitalization of manufacturing empowers enterprises’ low-emission development pathways. Building on this foundation, this paper reviews and synthesizes the existing scholarly research and conducts investigations and interviews with experts, scholars, business executives, and industry professionals engaged in intelligent manufacturing, cultural and creative product production, and digital transformation within manufacturing enterprises. Additionally, it integrates the authors’ extensive theoretical research and practical experience in the digitalization of manufacturing, explores the pathways, and assesses the effectiveness of different transformation approaches. Determining effective carbon reduction strategies across the full spectrum of digital-intelligent transformation in manufacturing enterprises—including product research and development, production, sales, and service—is crucial for steering enterprises toward a sustainable and eco-friendly development pathway.

This study makes two primary contributions: (1) It addresses a theoretical gap by examining the integration of digital transformation and sustainable development within manufacturing enterprises; and (2) It clarifies how digital transformation influences green, low-carbon, and sustainable development. The findings provide actionable insights for manufacturing enterprises to optimize their digital transformation processes, enhancing their contributions to green and sustainable development.

2. Digital Intelligence Empowering Green Low-Carbon Development Efficiency Measurement Index System and Measurement Model Construction

The digitalization of manufacturing enterprises can promote green and carbon-efficient progress through multiple channels, aspects, and pathways; however, the effectiveness of different pathways in advancing sustainable green development varies [26,27]. Investigating the underlying mechanisms of enterprise digital reinvention facilitates the advancement of sustainable low-emission development. Only under such conditions can the efficacy of each pathway be assessed, enabling the identification of those that yield the highest degree of optimization and impact [28,29].

2.1. Building an Index System to Measure the Effectiveness of Digital Intelligence and Empowerment of Green and Low-Carbon Development

In constructing a measurement index system for the effectiveness of digital transformation in enabling green and low-carbon development, methods such as literature reviews, brainstorming, and inductive summarization were primarily employed.

To begin with, an extensive review and synthesis of the existing literature on enterprise digitalization, digital-intelligent transformation, and environmentally sustainable, low-carbon development were conducted to identify well-established measurement indicators derived from prior research. Lin and Xie (2024) analyzed how digital transformation influences green innovation efficiency, concluding that technological advancement driven by digitalization enhances production efficiency, reduces waste, and improves energy structures, fostering low-carbon development [30]. Zhao et al. (2024) examined the role of digital transformation in enterprise pollution reduction, demonstrating that digital adoption enhances corporate governance, promotes clean production technologies, and supports long-term carbon neutrality strategies [31]. Chiarini et al. (2021) pointed out that enterprise digital-intelligent transformation expands the application scenarios of intelligent technologies [32]. By introducing cleaner and more intelligent production equipment and smart environmental monitoring devices, enterprises can achieve predictive maintenance, improve energy efficiency, and reduce carbon emissions [33]. Zhang et al. (2024) explored the relationship between enterprise digital transformation and carbon emissions, demonstrating that digitalization enables firms to transition to green business models, improve energy efficiency, and integrate sustainability strategies into industrial processes [34]. Lee et al. (2018) argued that digital-intelligent transformation helps eliminate information barriers between enterprises and financial systems, increasing financial institutions’ trust in enterprises and their willingness to provide funding for green and sustainable development [35]. Si et al. (2024) investigated how the digital economy drives green transformation in manufacturing enterprises, concluding that digitalization fosters proactive environmental initiatives and supports the adoption of clean technologies rather than merely responding to regulatory pressures [36]. From the existing research, key indicators for measuring the impact of digitally intelligent transformation in enhancing eco-friendly and low-carbon development include intelligent production substitution, implementation of intelligent scientific management, strengthening internal control to establish intelligent information disclosure platforms, optimizing energy structures, advancing the value chain to higher-end segments, reshaping business processes through intelligentization, environmental regulation, and increasing investment in intelligent environmental protection [37,38,39,40,41].

Secondly, the brainstorming method was employed to collect measurement indicators for the effectiveness of digital transition in facilitating low-carbon development. This brainstorming session was conducted with a panel of eight members, including four experts and scholars with senior professional titles and doctoral degrees in fields such as the digital economy, intelligent manufacturing of cultural and creative products, artificial intelligence, and enterprise innovation. These experts were affiliated with prestigious institutions, including Northwest University and Xi’an University of Architecture and Technology. The group also included four members from large-scale manufacturing enterprises, such as Fast Gear, Shaanxi Cultural Industry Investment Group Co., Ltd., Xi’an Tang West Market Culture Investment Group Co., Ltd., Shaangu Power, and XAC, comprising management personnel and frontline employees engaged in digitalization and intelligent manufacturing operations. The group shared their opinions and suggestions on the effect of enterprise digital-intelligent transition on sustainable and low-carbon development, focusing on relevant measurement indicators and pathways. The conclusions derived from the brainstorming sessions included the following:

• Digital technologies enable more precise material usage and production, reducing costs and emissions.
• Manufacturing enterprises are adopting more flexible and flattened organizational structures, improving responsiveness to market changes, reducing information asymmetry, and enhancing production efficiency [42].
• Digital transformation facilitates industrial upgrading, transitioning traditional manufacturing sectors toward digital service solution design.
• Digital transformation promotes refined management practices, reducing waste and saving costs.
• Integrating digital transformation with increased investment in green technology R&D and environmental protection directly enhances enterprises’ green development performance.
• Digital transformation empowers manufacturing enterprises’ products with higher value, improving quality and efficiency, thereby advancing sustainable development.

Drawing upon the findings from the literature review and the insights offered by the experts during the brainstorming sessions [43,44], the indicators used to assess the effectiveness of digital transformation in promoting green, sustainable, and low-carbon development can be classified into nine primary dimensions: intelligent production assessment, intelligent scientific management, intelligent production and manufacturing, upgrading manufacturing to intelligent services, empowering products with higher value, intelligent restructuring of business processes, investment in green innovative technologies, development of intelligent information disclosure platforms, and increased investment in smart environmental protection. This particular information is presented in

Table 1. The index system for measuring the effectiveness of digital transformation in promoting green and low-carbon development in manufacturing enterprises.

Measurement Objective	Measurement Indicators	Explanation of Indicator Connotations
The Effectiveness of Digital Transformation in Advancing Green and Low-Carbon Development	Intelligent Production Assessment (IPA)	Exemplifies Precision Production: This includes intelligent and precise calculations of required raw materials, production capacity, and material utilization during production. Through intelligent precision assessments, waste or excess raw material consumption is minimized.
	Intelligent Scientific Management (ISM)	Applies AI and Large Data Models: By embedding advanced technologies, including artificial intelligence and large-scale data models, into the management frameworks of manufacturing enterprises, intelligent information management systems can achieve lean, scientific management, resulting in energy conservation and emission reduction.
	Intelligent Production and Manufacturing (IPM)	Replaces Traditional Manufacturing Processes: Digital equipment and intelligent production technologies substitute traditional production methods. The use of AI and robotics enhances production efficiency while reducing pollutant emissions.
	Upgrading Manufacturing to Intelligent Services (UMIS)	Transforms Production into Intelligent Services: As the digital economy develops, traditional product manufacturing evolves to include intelligent service solutions, such as intelligent fault alerts and remote intelligent maintenance guidance. This transition supports carbon reduction, emission control, and efficient, green, sustainable development.
	Empowering Products with Higher Value (EPHV)	Empower Products with Higher Value: Intelligent technologies enable products to transition from low-end traditional manufacturing to high-end value chain production. This shift enhances product functionality and quality without significantly increasing production costs, achieving carbon reduction, emission control, and quality improvement.
	Intelligent Restructuring of Business Processes (IRBP)	Reconfigures Traditional Production Processes: Traditional manufacturing workflows undergo intelligent upgrades or restructuring. By optimizing and consolidating processes, establishing intelligent production platforms, and enhancing efficiency, pollution is reduced, resources are conserved, and sustainable development is achieved.
	Investment in innovative green technologies (IIGT)	Increases Investment in Green Innovation Technologies: Manufacturing enterprises amplify their R&D efforts in green innovative technologies. Applying AI and big data to input–output ratio calculations ensures higher innovation returns with increased resource allocation, resulting in cost reduction, efficiency gains, and carbon reduction.
	Development of Intelligent Information Disclosure Platforms (DIIDP)	Breaks Information Silos via Intelligent Platforms: Smart platforms enhance transparency in production and financial information, ensuring symmetry between the enterprise and external stakeholders. This boosts management efficiency and facilitates high-quality, sustainable development.
	Increasing Investment in Smart Environmental Protection (IISEP)	Reduces Pollutant Emissions through Smart Technologies: Intelligent technologies minimize pollutants in the manufacturing process. For instance, intelligent pollutant monitoring devices and innovative wastewater purification systems ensure compliance with emission standards, enhancing the enterprise’s low-carbon performance.

.

2.2. Construction of the Effectiveness Measurement Model

In constructing an efficiency measurement model for the digital-intelligent transformation of manufacturing aimed at promoting sustainable and low-carbon development, it is essential to employ a scientifically rigorous and logical evaluation method. The selection of an appropriate evaluation method directly influences the accuracy and reliability of efficiency assessments, ensuring that the transformation process effectively contributes to green and low-carbon objectives.

2.2.1. Comparison and Selection of Construction Methods of Measurement Model

Generally speaking, the methods for comprehensive evaluation of measurement indicators include the Analytic Hierarchy Process, Delphi method, fuzzy gray matter-element method, fuzzy normalized multi-attribute boundary approximate area comparison method, support vector regression, entropy weight TOPSIS (Technique for order preference by similarity to an ideal solution) evaluation method, etc. The advantages and disadvantages of different methods and their applicable scope and conditions are also different, specifically as follows:

The Analytic Hierarchy Process (AHP), proposed by American operations researcher Thomas Saaty in the 1970s, is a systematic and hierarchical multi-criteria decision-making method. The primary advantages of AHP include its simplicity and low data requirements; however, its limitations are evident in its strong subjectivity, increased workload when multiple factors are involved, high proportion of qualitative elements, and difficulty in handling dynamically changing decision-making environments [45]. The Delphi method, introduced by the RAND Corporation in the 1950s, is a structured expert consultation decision-making approach. This method is advantageous for addressing complex problems and integrating diverse expert perspectives. However, it is time-consuming, costly, heavily reliant on expert input, and characterized by a high degree of subjectivity [46]. The fuzzy gray matter-element method is a decision-making approach designed to handle complex macro-system problems involving multiple factors, uncertainties, and incomplete information. It integrates fuzzy set theory and gray system theory, allowing for the simultaneous processing of quantitative data, qualitative descriptions, and incomplete information. This makes it suitable for multi-dimensional evaluation in complex systems. However, it also has certain limitations, such as high computational complexity, strong dependence on expert experience, and a relatively high application threshold, which may restrict its broader implementation [47,48]. The fuzzy normalized multi-attribute boundary approximate area comparison method is employed in multi-attribute decision analysis, primarily by incorporating potential loss and gain values for compensation. This method is particularly useful for dealing with fuzzy information and uncertainty-related problems. However, it has stringent data requirements, involves high computational complexity, and produces results that are relatively difficult to interpret [49]. Support Vector Regression (SVR) leverages kernel functions to handle nonlinear relationships and is particularly suitable for high-dimensional data and small-sample scenarios. However, it has high computational complexity, is highly sensitive to parameter selection, and exhibits challenges in intuitively interpreting nonlinear kernel function models [50,51].

The entropy-weighted TOPSIS evaluation method integrates entropy weighting with the TOPSIS approach [52], balancing data objectivity with decision-making logic. It is well suited for multi-criteria evaluation, providing scientifically rigorous and easily interpretable results, and has been widely applied in the fields of economics and management. Its shortcomings are as follows: First, the correlation of evaluation indicators has not been effectively dealt with, and the entropy weight method assumes that the indicators are independent. However, in actual evaluation, there is sometimes a correlation between indicators. Therefore, it is essential to eliminate correlations using SPSS 22.0 (Statistical Product and Service Solutions) software through descriptive statistical analysis and correlation analysis. Second, it is susceptible to outliers and data quality in the evaluation index. The occurrence of outliers will distort the dispersion of data, leading to the distortion of weights and standardization results. Therefore, during the data cleaning and analysis phase, the reliability and validity of the collected data are evaluated to identify and address any anomalies, ensuring the overall integrity of the data. Assessing the efficiency of digital-intelligent transformation in manufacturing enterprises to foster sustainable and low-carbon development presents a multi-criteria challenge. The data collected through investigations and comprehensive evaluations of the pathway indicators in digital-intelligent transformation aimed at advancing sustainable and low-carbon development are inherently discontinuous. Due to this discontinuity, the SVR method is unsuitable for assessing the relative contribution of various pathway indicators in manufacturing enterprises’ digital-intelligent transformation toward achieving the overarching sustainability objectives. The conclusion is required to be general and instructive. On the whole, the entropy weight TOPSIS evaluation method is more suitable. The conclusions are more objective and scientifically robust by integrating statistical analysis with indicator correlation assessment and abnormal data processing. This approach provides scientifically grounded guidance and broad applicability. It offers manufacturing enterprises valuable insights into selecting more effective pathways to advance sustainable and low-carbon development during their digital-intelligent transformation process.

2.2.2. Entropy Weight-Based TOPSIS Method

The Entropy Weight-based TOPSIS method is a multi-criteria decision-making approach that combines the entropy weight method with the TOPSIS technique. This approach uses the entropy weight method to assign weights to the indices, which are then applied in the TOPSIS technique to identify the optimal solution [53]. The entropy weight method, grounded in information entropy theory, evaluates the significance of indicators by quantifying the uncertainty associated with the information they provide. A lower information content corresponds to more significant uncertainty and a higher entropy value, whereas a higher information content leads to reduced uncertainty and a lower entropy value. The TOPSIS technique, a multi-criteria decision analysis method, begins by normalizing the original data matrix. It then identifies the optimal and worst alternatives from a finite set of options. The relative distances between each evaluation object and the optimal and worst alternatives are calculated to establish their comparative ranking. The primary criterion for performance evaluation is the relative closeness of each evaluation object to the optimal solution [54].

The principle and main steps of entropy weight TOPSIS evaluation method are as follows [55,56]:

(1)

Standardized data processing. First, the trend is the same. The comprehensive evaluation indicators include high-excellent, intermediate, and low-excellent indicators. In order to ensure that all evaluation indicators are in the same direction, the low-excellent indicators are usually highly optimized, and the difference method $\left(1-x\right)$ is used to use the reciprocal method $\left(1/x\right)$ for the absolute low-excellent indicators, and the intermediate indicators are highly optimized through the best values. Second, the index is dimensionless. It mainly eliminates the influence of index measurement unit on evaluation and normalizes the measured value of the index without dimension. Let ${\left(X_{ij}\right)}_{n\times m}$ be the indicator matrix after the same trend and ${\left(P_{ij}\right)}_{n\times m}$ be the normalized data matrix, then $P_{ij}=X_{ij}/\sqrt{{\displaystyle \sum }_{i=1}^n{X}_{ij}^2}$.

(2)

Determine the weight. The first step involves determining the proportion of the $i$ factor within the context of the $j$ index $e_j=-k{\displaystyle \sum }_{i=1}^n{P}_{ij}In\left(P_{ij}\right)$. The second step consists of calculating the entropy value for the j-th index, denoted as $k=1/In\left(n\right)>0$. The third step is to evaluate the redundancy of the information entropy and compute the weight for each index, represented by $d_i=1-e_i$, $w_j=d_i/{{\displaystyle \sum }}_{j=1}^m{d}_j$.

(3)

Identify the optimal and worst schemes from the available alternatives. If the original data are unified into high-quality indicators after the same trend, the optimal scheme is $X^{+}=\left(X_1^{+},X_2^{+},K,X_m^{+}\right)$; the worst scheme is $X^{-}=\left(X_1^{-},X_2^{-},K,X_m^{-}\right)$.

(4)

Compute the weighted Euclidean distance for each evaluation index. The optimal scheme is denoted as $D_i^{+}=\sqrt{{\displaystyle \sum }_{k=1}^q{\left[X_i\left(k\right)-X_0^{+}\left(k\right)\right]}^2}$, and the worst scheme is $D_i^{-}=\sqrt{{\displaystyle \sum }_{k=1}^q{\left[X_i\left(k\right)-X_0^{-}\left(k\right)\right]}^2}$.

(5)

Assess the extent to which each evaluation index approaches the optimal solution, denoted as $D_i^{*}=D_i^{-}/D_i^{-}+D_i^{+}$.

2.2.3. Entropy Weight TOPSIS Evaluation Model Construction

In constructing an efficiency measurement model for the digital-intelligent transformation of manufacturing enterprises aimed at promoting sustainable and low-carbon development, the entropy weight TOPSIS evaluation model was utilized. This study explored how digital-intelligent transformation contributes to advancing sustainable and low-carbon development paths for enterprises by conducting surveys and collecting data from experts, scholars, managers of manufacturing enterprises, and professionals in digital-intelligent business operations. The analysis identified the most effective transformation strategies for achieving sustainability goals [57,58,59]. In line with the principles of the entropy weight TOPSIS method, this study evaluated the efficiency of digital-intelligent transformation in manufacturing enterprises for driving sustainable development and developed the corresponding measurement model.

$$X_i\left(k\right)=X_i^{’}\left(k\right)/{\displaystyle \frac{1}{p}}{\displaystyle \sum }_{i=1}^p{X}_i^{’}\left(k\right)\left(k=1,2,\cdots ,q\right)$$

(1)

The comprehensive measurement evaluation model represents the object being measured or evaluated with measurement indicators. In this study, the indicators included nine items: intelligent production assessment, intelligent scientific management, intelligent production and manufacturing, upgrading manufacturing to intelligent services, empowering products with higher value, intelligent restructuring of business processes, investment in green innovative technologies, development of intelligent information disclosure platforms, and increasing investment in smart environmental protection. These form the evaluation indicator set, creating a data matrix of the indicators. Given that all selected measurement indicators are positive, the collected data can be standardized using the mean normalization method. This approach involves computing the ratio of each sample data point to the mean of all corresponding indicator values, as expressed in Equation (1), thereby generating a standardized matrix.

$$A=\left(X_1,X_2,L,X_p\right)=\left[\begin{array}{c}\begin{array}{c}{X}_1\left(1\right),X_1\left(2\right),\cdots ,X_1\left(q\right)\\ X_2\left(1\right),X_2\left(2\right),\cdots ,X_2\left(q\right)\end{array}\\ \begin{array}{c}\cdots ,\cdots ,\cdots ,\cdots \\ X_p\left(1\right),X_p\left(2\right),\cdots ,X_p\left(q\right)\end{array}\end{array}\right]$$

In Equation (1), the data for the i-th sample is: $X_i=\left(X_i\left(1\right),X_i\left(2\right),\cdots ,X_i\left(q\right)\right)\left(i=1,2,\cdots ,p\right)$.

(1) Entropy Weight Calculation for the Standardized Matrix

First, the entropy weight for each indicator is computed [60]. Based on the standardized data, the indicator proportions $B_i\left(k\right)$, entropy values $e\left(k\right)$, and indicator variation coefficients $g\left(k\right)$ are computed. Finally, the weights for each indicator $w\left(k\right)$ are determined (Equations (2)–(5)).

$$B_i\left(k\right)=X_i\left(k\right)/{\displaystyle \sum }_1^p{X}_i\left(k\right)$$

(2)

$$e\left(k\right)={\displaystyle \frac{1}{InP}}{\displaystyle \sum }_1^P{B}_i\left(k\right)In{B}_i\left(k\right),0\le e\left(k\right)\le 1$$

(3)

$$g\left(k\right)=1-e\left(k\right),0\le g\left(k\right)\le 1$$

(4)

$$W\left(k\right)=g\left(k\right)/{{\displaystyle \sum }}_{i=1}^{p}g\left(k\right),0\le w\left(k\right)\le 1,w\left(1\right)+w\left(2\right)+\cdots +w\left(q\right)=1$$

(5)

Secondly, compute the weight set $C$ for the standardized matrix (Equation (6)):

$$C=A\times W$$

(6)

(2) Relative Proximity Calculation Using Euclidean Distance

Initially, determine the positive ideal solution (PIS) and the negative ideal solution (NIS). Precisely determine the optimal and worst values for each of the nine indicators among the measurement objects. These values are then used to formulate the PIS and NIS, as shown below:

Positive ideal solution: $X_0^{+}=\left(X_0^{+}\left(1\right),X_0^{+}\left(2\right),\cdots ,X_0^{+}\left(k\right),\cdots ,X_0^{+}\left(n\right)\right),\left(k=1,2,\cdots ,q\right)$

Negative ideal solution: $X_0^{-}=\left(X_0^{-}\left(1\right),X_0^{-}\left(2\right),\cdots ,X_0^{-}\left(k\right),\cdots ,X_0^{-}\left(n\right)\right),\left(k=1,2,\cdots ,q\right)$

Next, utilizing the weight set $B$ of the standardized matrix, compute the distance $D_i^{+}$ of the nine measurement indicators from the ideal solution (Equation (7)):

$$D_i^{+}=\sqrt{{\displaystyle \sum }_{k=1}^q{\left[X_i\left(k\right)-X_0^{+}\left(k\right)\right]}^2}$$

(7)

Calculate the distance $D_i^{-}$ from 9 measurement indicators to the negative ideal solution (Equation (8)):

$$D_i^{-}=\sqrt{{\displaystyle \sum }_{k=1}^q{\left[X_i\left(k\right)-X_0^{-}\left(k\right)\right]}^2}$$

(8)

Finally, the relative proximity is computed using the Euclidean distance, as defined in Equation (9). This metric evaluates the contribution values linked to various pathways through which digital transformation in manufacturing enterprises supports sustainable and low-carbon development. The results are subsequently ranked in descending order, serving as a basis for further outcomes analysis.

$$D_i^{*}=D_i^{-}/D_i^{-}+D_i^{+}$$

(9)

3. Data Source and Test Analysis

When assessing the effectiveness of digital transformation in facilitating sustainable and low-carbon development in manufacturing enterprises, the primary data sources included validated survey questionnaires and structured interviews. Surveys and interviews were conducted with experts from Double First-Class universities, key provincial universities, intelligent manufacturing research institutes, think tank experts specializing in intelligent manufacturing and digital transformation, enterprise managers, and practitioners in fields such as intelligent manufacturing and product production, digital transformation, and the digital economy. The primary methods for data collection included distributing questionnaires via email, WeChat platforms, alum networks, and in-depth enterprise interviews. The nine indicators were assessed using a 7-point Likert scale [61,62,63].

3.1. Design and Refinement of the Questionnaire

To collect data for evaluating the effectiveness of digital transformation in promoting sustainable and low-carbon development in manufacturing enterprises, a survey questionnaire was created using the nine effectiveness measurement indicators. These indicators were transformed into related questions. The questionnaire was structured into three primary sections, as follows:

• Basic Information: Collects demographic and professional details of the respondents.
• Core Content: The survey focuses on the nine measurement indicators and employed objective questions rated on a 7-point Likert scale. This scale, ranging from 1 to 7, quantifies how digital transformation contributes to sustainable and low-carbon development in manufacturing enterprises. Specifically, a rating of 1 signifies no contribution, 2 represents a very weak contribution, 3 indicates a weak contribution, 4 denotes a neutral stance, 5 reflects a substantial contribution, 6 signifies an influential contribution, and 7 corresponds to a powerful contribution.
• Open-Ended Questions: This section captures respondents’ suggestions on more effective pathways for digital transformation as a key driver for advancing sustainable and low-carbon development.

To verify the validity and reliability of the survey data [64], a preliminary survey was administered to 10 experts and industry leaders specializing in intelligent manufacturing. Feedback from the pilot responses was used to refine the questionnaire, resulting in a robust and reliable final version.

3.2. Distribution and Collection of the Survey Questionnaire

The formal survey began in early September 2024 and concluded in November, spanning nearly three months. A total of 300 questionnaires were disseminated, of which 288 were recovered. After excluding invalid or incomplete responses, 265 valid questionnaires remained, resulting in an effective response rate of 88.33%. The target respondents comprised scholars holding doctoral degrees and experts with senior professional titles from nine universities, including Northwest University, Northwestern Polytechnical University, Kunming University of Science and Technology, and Anhui University, specializing in fields such as digital economy, cultural creativity, artificial intelligence, intelligent manufacturing, and mechanical engineering. The survey also involved managers and frontline employees from over 20 large-scale manufacturing enterprises, including Fast Gear, Shaanxi Cultural Industry Investment Group Co., Ltd., Shaangu Power, XAC, and Shaanxi Automobile Group, working in digitalization and intelligent manufacturing.

3.3. Statistical Analysis of the Data

The survey and interview data were subjected to statistical analysis [43,65], focusing on reliability and validity tests. Outlier data were excluded to ensure that the data aligned with the research questions and supported the accuracy of the conclusions. Statistical analyses were performed using SPSS software (version 22), with the descriptive statistics shown in

Table 2. Descriptive statistics.

	N	Min	Max	Mean	SD
IPA	265	2	7	6.1333	0.96956
ISM	265	3	7	5.9500	1.04399
IPM	265	2	7	6.0417	1.07215
UMIS	265	2	7	5.8167	1.19511
EPHV	265	2	7	5.8500	1.08194
IRBP	265	2	7	5.9250	1.09362
IIGT	265	4	7	6.1083	0.89627
DIIDP	265	2	7	5.8917	1.11367
IISEP	265	2	7	5.9750	1.04891
Valid N	265

.

The statistical analysis results shown in Table 2 showed minimal differences in the standard deviations of the survey data, indicating a low overall dispersion of the data [66]. The data collected on intelligent production evaluation, intelligent scientific management, intelligent manufacturing, and the transition of manufacturing operations to intelligent services effectively captured the efficacy of digital transformation in advancing sustainable development in manufacturing. This evidence provides robust support for subsequent research.

To further examine the interrelationships among the indicators within the efficiency measurement index system—designed to facilitate low-carbon development through the digital-intelligent transformation of manufacturing—and to ensure that the gathered statistical data accurately reflected the efficiency of such transformative processes, the reliability and validity of the correlations between indicators and statistical data were evaluated using factor analysis [67,68,69]. Utilizing SPSS (version 22) for social science statistical analysis, the outcomes of the correlation analysis among the indicators are displayed in

Table 3. Correlation matrix.

Correlation Coefficient	IPA	ISM	IPM	UMIS	EPHV	IRBP	IIGT	DIIDP	IISEP
IPA	1.000	0.326	0.415	0.257	0.000	0.453	0.000	0.102	0.000
ISM	0.326	1.000	0.201	0.352	0.346	0.475	0.446	0.364	0.344
IPM	0.415	0.201	1.000	0.265	0.418	0.210	0.441	0.440	0.000
UMIS	0.257	0.352	0.265	1.000	0.325	0.423	0.450	0.484	0.392
EPHV	0.000	0.346	0.418	0.325	1.000	0.402	0.476	0.475	0.000
IRBP	0.453	0.475	0.210	0.423	0.402	1.000	0.235	0.365	0.475
IIGT	0.000	0.446	0.441	0.450	0.476	0.235	1.000	0.465	0.328
DIIDP	0.102	0.364	0.440	0.484	0.475	0.365	0.465	1.000	0.398
IISEP	0.000	0.344	0.000	0.392	0.000	0.475	0.328	0.398	1.000

.

The correlation matrix values presented in Table 3 illustrate the relationships among the indicators within the efficiency measurement index system designed to evaluate the digital-intelligent transformation of manufacturing enterprises and its contribution to green and low-carbon development. Analysis of the correlation coefficients showed that the nine indicators are either independent or exhibit minimal correlation. This finding validates the appropriateness of applying the entropy weight TOPSIS evaluation method to assess the effectiveness of digital-intelligent transformation in supporting green and low-carbon development. Consequently, the constructed index system provides a robust basis for measuring the pathway weights of digital-intelligent transformation in manufacturing enterprises aimed at fostering green and low-carbon development while also being suitable for factor analysis to confirm the reliability and validity of the collected data [70].

3.3.1. Reliability Test

Cronbach’s method was applied to assess the reliability of the statistical data [71]. Utilizing SPSS (version 22), the Cronbach’s alpha coefficient was computed, as presented in

Table 4. Reliability statistics.

Cronbach’α	Cronbach’α (Standardized)	Number of Items
0.900	0.900	9

.

As presented in Table 4, the Cronbach’s alpha coefficient for the statistical data assessing the effectiveness of digital transformation in advancing green and low-carbon development in manufacturing enterprises was 0.900. This value signifies a high level of reliability in the dataset [72,73].

3.3.2. Validity Test

Exploratory factor analysis was conducted using SPSS for statistical analysis in the social sciences to assess the validity of the data. The analysis resulted in a Kaiser–Meyer–Olkin (KMO) value of 0.873 and a Bartlett’s test of sphericity value of 574.561, as shown in

Table 5. KMO value and Bartlett’s test.

KMO	Bartlett’s Test	Df	Significance
0.873	574.561	36	0.000

. These findings demonstrate that the measurement indicator system possesses strong structural validity and offers robust explanatory power for assessing the effectiveness of digital transformation in advancing green and low-carbon development in manufacturing enterprises [74,75].

Drawing from the outcomes of the reliability and validity assessments, the data obtained through surveys and interviews for the nine measurement indicators—intelligent production assessment, intelligent scientific management, intelligent production and manufacturing, upgrading manufacturing to intelligent services, empowering products with higher value, intelligent restructuring of business processes, investment in green innovative technologies, development of intelligent information disclosure platforms, and increasing investment in smart environmental protection—demonstrated good reliability and validity. The data provided strong empirical support for assessing the effectiveness of digital transformation in advancing green and low-carbon development in manufacturing enterprises.

4. Effectiveness Measurement and Result Analysis

Employing the entropy-weighted TOPSIS comprehensive evaluation method, the contribution of the effectiveness indicators for digital transformation in advancing low-carbon development in manufacturing was quantified using valid data from the surveys and interviews.

4.1. Effectiveness Measurement

First, the data were standardized, and the entropy weights for the standardized data, ranging from 1 to 7, were calculated. Next, the PIS and NIS were determined, as follows:

Positive ideal solution: $X_0^{+}=\left(X_0^{+}\left(1\right),X_0^{+}\left(2\right),\cdots ,X_0^{+}\left(k\right),\cdots ,X_0^{+}\left(n\right)\right),\left(k=1,2,\cdots ,7\right)$

Negative ideal solution: $X_0^{-}=\left(X_0^{-}\left(1\right),X_0^{-}\left(2\right),\cdots ,X_0^{-}\left(k\right),\cdots ,X_0^{-}\left(n\right)\right),\left(k=1,2,\cdots ,7\right)$

Next, using Equation (7), the distances of the nine measurement indicators from the ideal solution were calculated based on the weight set derived from the standardized matrix. Applying Equation (8), the distances from the NIS were calculated. Subsequently, Equation (9) was utilized to compute the relative proximity based on Euclidean distance, providing a comprehensive evaluation of the contribution of each measurement indicator to the effectiveness of digital transformation in facilitating low-carbon development in manufacturing. The indicators were then ranked in descending order to analyze the final results. This approach comprehensively reflects the contribution levels of the nine measurement indicators: intelligent production assessment, intelligent scientific management, intelligent production, and manufacturing, upgrading manufacturing to intelligent services, empowering products with higher value, intelligent restructuring of business processes, investment in green innovative technologies, development of intelligent information disclosure platforms, and increasing investment in smart environmental protection. The entropy values of these contributions are presented in

Table 6. Entropy values of measurement indicators for assessing the effectiveness of digital transformation in facilitating green and low-carbon development in manufacturing enterprises.

Evaluation Indicators	Information Provider	Measurement Results	Weight Ranking
IPA	0.99311	0.06384	9
ISM	0.98315	0.15614	2
IPM	0.99109	0.08249	7
UMIS	0.98735	0.11723	3
EPHV	0.99033	0.08966	6
IRBP	0.99026	0.09024	5
IIGT	0.97569	0.22523	1
DIIDP	0.98974	0.09505	4
IISEP	0.99135	0.08013	8

.

To ensure the accuracy and reliability of the measurement results, a cumulative weight-based random selection method, derived from the stochastic weight recognition approach, was employed for verification and in-depth analysis. Typically, stochastic weight calculation is applied to randomly select among multiple options based on their respective weights. In this context, weights represent both the proportion of workload distribution and the degree of influence on overall capability. Common approaches to stochastic weight calculation include uniform distribution-based random weighting and cumulative weight-based random selection [76,77]. Given the specific focus of this study on measuring the effectiveness of various pathways through which digital intelligence enables green and low-carbon development in manufacturing enterprises, the cumulative weight-based random selection method was deemed the most appropriate for analysis.

In Table 6, the information entropy of the measurement indicators for the effectiveness of digital intelligence-driven green and low-carbon transformation in manufacturing enterprises serves as a metric to assess the degree of information concentration. It reflects the maximum amount of information when the sum of probabilities for all possible values in a random variable equals one. The calculation formula is given as $H\left(x\right)=-\sum P\left(x\right)·lo{g}_{2}P\left(x\right)$, where $P\left(x\right)$ represents the probability of the i-th indicator occurring across all samples. The differentiation coefficient of each indicator is derived using $1-H\left(x\right)$, while the weight of each measurement indicator is determined by applying the formula $w_i={\displaystyle \frac{\sum P\left(x\right)·H\left(x\right)}{H\left(D\right)}}$. In this equation, $w_i$ denotes the weight, $P\left(x\right)$ signifies the probability of the i-th indicator appearing in all samples, and $H\left(D\right)$ represents the total information entropy of the dataset. The results of the differentiation coefficient and the weight calculation for the measurement indicators are presented in

Table 7. Difference coefficient and weight of efficiency measurement indicators.

Evaluating Indicator	Difference Coefficient	Weight	Sum of Weights
IPA	0.00689	0.06384	1
ISM	0.01685	0.15614
IPM	0.00890	0.08249
UMIS	0.01265	0.11723
EPHV	0.00967	0.08966
IRBP	0.00974	0.09024
IIGT	0.02431	0.22523
DIIDP	0.01026	0.09505
IISEP	0.00865	0.08013

.

Subsequently, a random number within the range of [0,1] was generated using Python (version 3.9.7) programming. The randomly generated value, 0.09, was then employed in conjunction with the entropy-weighted TOPSIS comprehensive evaluation method to compute the weights of the measurement indicators for the effectiveness of digital intelligence-driven green and low-carbon transformation in manufacturing enterprises. The selected random number (0.09) was deemed appropriate, as it effectively differentiates the contributions of various pathway effectiveness measurement indicators.

Finally, the corresponding indicators were determined based on the position of the randomly generated number 0.09. If an indicator’s effectiveness measurement value is greater than or equal to 0.09, it signifies a relatively high contribution to effectiveness and qualifies as a selectable random variable. Conversely, if the effectiveness measurement value falls below 0.09, the indicator’s contribution is relatively low, suggesting a need for careful consideration or potential exclusion. According to the measurement results, five indicators—green innovation technology investment, intelligent scientific management, intelligent service for manufacturing upgrades, clever information disclosure platform construction, and intelligent business process restructuring—demonstrated effectiveness measurement values exceeding 0.09, making them viable random variables for selection. Meanwhile, four indicators—enhancing product value through digital empowerment, intelligent production and manufacturing, increasing investment in intelligent environmental protection, and intelligent production evaluation—had effectiveness measurement values below 0.09. However, among them, the first three indicators had values close to 0.08, warranting careful consideration rather than outright exclusion. These findings provided a robust foundation for analyzing the effectiveness measurement of digital intelligence-driven green and low-carbon transformation in manufacturing enterprises, ensuring the reliability and validity of the analysis results.

4.2. Robustness Verification of Measurement Results

The primary approach involved modifying the sample range when evaluating the robustness of the path index weights associated with the green and low-carbon development of manufacturing enterprises driven by digital intelligence. Specifically, the variation in weight values and ranking was examined as the sample size of the collected data fluctuated to assess the stability of the measurement results. Weight calculations for the green and low-carbon development path indicators were initially conducted using 265 samples. Subsequently, the data collection was extended using the same research methodology, incorporating 15 academic experts. Moreover, the sample was further expanded to include 35 enterprise managers and professionals specializing in intelligent manufacturing, digital transformation, and the digital economy from leading corporations, such as Siemens AG (Munich, Germany), Bavarian Engine Factory Co., Ltd. (Munich, Germany), Mitsubishi Group (Tokyo, Japan), and General Electric Company (Boston, MA, USA). As a result, a total of 336 questionnaires were obtained. After excluding invalid responses, 312 valid samples were selected for analysis, yielding a reliable response rate of 89.14%. The dataset underwent descriptive statistical analysis, as well as reliability and validity testing, using the SPSS software. Any anomalies detected were removed, and the entropy weight TOPSIS method was used to calculate the distance to the NIS and the relative closeness based on the Euclidean distance. When ranked in descending order, the expanded sample’s weight calculations aligned with the original sample’s results. This consistency demonstrates the entropy-weighted TOPSIS approach’s robustness, confirming the research findings’ objectivity and scientific validity.

4.3. Analysis of Effectiveness Measurement Results

The measurement results were further subjected to statistical analysis, and the entropy values, along with the weight rankings of the efficiency measurement indicators for low-carbon development, facilitated by the transformation of manufacturing enterprises through digital and intelligent technologies. (as presented in Table 6) were visually represented through statistical graphs, as shown in Figure 2.

Based on the measurement results and statistical analysis, it is evident that the contributions of the nine indicators to the efficiency of green and low-carbon development in manufacturing enterprises differ. Among these, investment in green innovation technology is particularly significant, ranking first with the highest contribution rate of 22.5%, surpassing the second-ranked intelligent scientific management by 7%. The contribution of intelligent scientific management, at 15.6%, is 4% higher than that of the intelligent service pathway. This indicates that digital-intelligent transformation is most effective in improving green and low-carbon performance through increased investment in green innovation and enhanced scientific management. Additionally, the transformation process advances green and low-carbon performance through manufacturing upgrades and intelligent services, with a contribution rate of 11.7%, highlighting their role in fostering sustainable development.

The establishment of intelligent information disclosure platforms, restructuring business processes, empowering products with higher value, intelligent production, and increased investment in smart environmental protection ranked fourth to eighth in contributing to green and low-carbon sustainable development, with contribution values between 8% and 9%. This suggests that these pathways exert a weaker influence on enhancing low-carbon performance and high-quality, sustainable development. The lower effectiveness may be due to higher costs, pollutants, and carbon emissions during the digital transformation, offsetting post-transformation reductions. As a result, the input–output ratio was suboptimal, and the expected outcomes were not met. Ranking ninth was intelligent production assessment, which focuses on optimizing material utilization to minimize waste. However, its contribution to sustainable development was limited, with a contribution rate of just 6.3%.

Drawing from prior research and the preceding analysis, it is clear that digital transformation is essential for promoting low-carbon development. This conclusion has been validated by enterprise managers, industry professionals, and scholars, leading to a shared consensus on its effectiveness. The digital-intelligent transformation in manufacturing promotes sustainable development through various pathways, including intelligent production measurement, scientific management, intelligent manufacturing, and upgrading manufacturing services. However, the effectiveness and contribution of different pathways to sustainable development vary, diverging from the existing literature’s conclusions. The disparity in efficiency contributions partially reflects the economic benefits of an enterprise’s input–output ratio. Specifically, while manufacturing enterprises allocate equivalent human, financial, and material resources toward enhancing green innovation performance in digital and intelligent transformation, the impact on resource-efficient performance differs depending on the chosen investment direction and pathway. For instance, allocating the same level of investment to accelerating research and development and applying green innovation technology significantly enhances enterprises’ green performance. In contrast, prioritizing product functionality and added value enhancement results in a comparatively lower contribution to carbon reduction and emissions mitigation. Additionally, if manufacturing enterprises aim to achieve an equivalent level of green performance improvement and emissions reduction through digital-intelligent transformation, the required investment varies depending on the chosen pathway. High-efficiency contribution pathways demand lower labor and capital input costs, whereas low-efficiency pathways necessitate more significant expenditures. Thus, enterprises should prioritize pathways with higher efficiency in facilitating green, sustainable, and low-carbon development throughout the digital-intelligent transformation. This can be achieved by increasing investment in green innovation technology to enhance the intelligence of management processes and by driving the transformation and upgrading of traditional manufacturing toward intelligent services. Such an approach will enable enterprises to achieve cost reduction, efficiency improvement, emissions reduction, and long-term green sustainability. The research findings provide valuable practical guidance and comparative insights into input–output economic benefits, helping manufacturing enterprises leverage digital innovations more effectively to support sustainable and low-carbon development.

5. Digital Intelligence Empowering Green and Low-Carbon Development Path Suggestions

Drawing on the prior analysis and the assessment outcomes regarding the effectiveness of digital transformation in advancing eco-friendly and low-carbon development, manufacturing enterprises can optimize the efficiency of their sustainable, low-carbon progress by utilizing the following strategic pathways throughout the digitalization process.

First, prioritize increased investment in the innovation, research, and practical implementation of green intelligent environmental protection technologies. Manufacturers can leverage big data technology to build an environmental protection cloud data center to advance digital transformation. This can be achieved through four layers: terminal aggregation, data management, intelligent optimization, and visualized management. The center would integrate and develop key technical services, such as intelligent monitoring, predictive analytics, risk assessment, and management strategies to assist in decision-making, thereby enabling intelligent environmental management at a technical level. By employing advanced artificial intelligence (AI) and big data technologies, enterprises can implement intelligent security systems, such as smart water quality monitors, air quality detectors, and dust monitoring devices, to more accurately monitor and identify pollution sources in production processes. These systems can also effectively predict and address environmental emergencies, enhancing the efficiency and precision of environmental protection efforts. Furthermore, AI technologies can be applied to the intelligent monitoring and fault detection of environmental protection equipment, systems, monitoring platforms, and pollution control mechanisms. This approach ensures that manufacturing processes achieve energy conservation, reduced emissions, and long-term sustainable green development.

Secondly, intelligent information systems and management platforms should be established to achieve manufacturing enterprises’ lean and scientific management processes. By incorporating advanced technologies, including artificial intelligence and large-scale data models, into production management practices, enterprises can build upon existing management modules (e.g., information management and production management) and adopt intelligent planning, execution, and control mechanisms supported by smart decision-making. This approach enables the intelligent allocation of enterprise resources and establishes a management system characterized by efficient integration of “machine elements” (hardware and software) and human–machine coordination. Promoting the intelligent development of systems such as Computer Integrated Manufacturing Systems, Enterprise Resource Planning, Supply Chain Management, and Customer Relationship Management is essential. By establishing intelligent information management systems, enterprises can implement lean and data-driven management, ultimately enhancing energy efficiency, reducing emissions, lowering costs, improving operational effectiveness, and fostering green, low-carbon, and sustainable development. By aligning with business needs, intelligent information management systems can enhance the integration of information flow and business flow, address issues of collaboration and integration among different information systems, and optimize processes while ensuring data integration and sharing. Comprehensive analysis of internal and external information needs and connections within the enterprise can guide the scientific and rational design of business architectures for manufacturing information systems, facilitating lean management. By enabling digital production equipment to be intelligently networked through wireless transmission, hardware devices can be interconnected, achieving digital and intelligent management of production and manufacturing processes.

Thirdly, harness digital and intelligent technologies to facilitate product upgrading and diversification while actively advancing the expansion of the digital service industry chain. With the advancement of the digital economy, traditional product manufacturing is transitioning to integrated solutions combining product production with intelligent service design, such as intelligent fault alerts and remote astute maintenance guidance. This shift supports carbon reduction, emission control, and high-efficiency green sustainable development. Research findings indicate that intelligent services are instrumental in markedly enhancing eco-efficient total factor productivity, serving as a fundamental catalyst for enterprises’ effective transition toward sustainability. Intelligent services represent a crucial development direction for advanced manufacturing, with broad application scenarios spanning the industrial internet, flexible customization, collaborative manufacturing, smart logistics, intelligent supply chains, and remote contactless operations and maintenance. Developing intelligent services in manufacturing enterprises primarily involves enabling intelligent service technologies, coordinating intelligent products with digital services, and applying intelligent service scenarios. By adopting intelligent service technologies, enterprises can enhance user insights, providing a clear understanding of user profiles, habits, and preferences. These insights can be integrated into product R&D and design, allowing for timely optimization. This approach effectively aligns production with demand, increases flexibility, improves resource utilization efficiency, and achieves high-efficiency, low-consumption production. Furthermore, applying intelligent technologies for data analysis and mining enhances insights into processes such as procurement and inventory, enabling precise management and promoting energy savings and consumption reduction. Lastly, intelligent service technologies support the establishment of ecological cycles within enterprises, encompassing reuse, remanufacturing, and recycling. This improves environmental performance and positions intelligent services as an intrinsic and practical driver of green transformation.