**4. The Fuzzy DEMATEL Method of Key Drivers**

The proposed model framework for analyzing the drivers of CER in China's automobile industry is shown in Figure 1. First of all, with the help of existing literature, expert opinions and industry managers' opinions, most of the drivers were collected. Second, a questionnaire containing the drivers of the five-point Likert scale was distributed to Chinese automobile companies. For the collected valid questionnaires, we averaged the survey results, discussed with experts, and finally identified and classified 16 common driving factors. Then, based on the TBL, a fuzzy direct relationship matrix was established, and the key drivers of CER in CAMI were analyzed using the DEMATEL program. Finally, the results were verified in a medium-sized automobile company through feedback from the automotive industry managers and comparison with existing literature.

**Figure 1.** The model framework for analyzing the drivers of CER in Chinese auto industry.

Since the drivers for the implementation of environmental responsibility in CAMI is a complex decision-making problem, it is a common method to use multicriteria decisionmaking (MCDM) [50,51] method or fuzzy analytic hierarchy process [52] to make decisions. DEMATEL, as one of the MCDM approaches, firstly used by The Battelle Memorial Institute at its Geneva Research Centre in 1973, is utilized as a solution method in this paper [53]. The DEMATEL method can visualize complex causal structures by establishing and analyzing structural models between complex factors. Furthermore, it can analyze the influence relationship between complex criteria and separate the factors into cause group and effect group in which the cause group affects the effect group thus reckoning the relative weights of criteria. In this paper, since the interaction between all the drivers of CER in CAMI is relatively complex, it is necessary to use DEMATEL to help us better understand the interaction between drivers.

Although DEMATEL is a good way to deal with complex decision-making problems, the degree of mutual influence between systems is usually ambiguous, which will make language information unsuitable for expression. In order to reduce uncertainty and increase accuracy, DEMATEL is combined with fuzzy logic proposed by Zadeh [54]. It is rather effective to measure the ambiguous concepts related to human's subjective judgments with fuzzy logic [55]. Therefore, this paper uses fuzzy DEMATEL with triangular fuzzy numbers to evaluate the driving factors of CER in CAMI.

A triangular fuzzy number can be defined as a triplet *A*- = (*l*, *m*, *u*), where *l*, *m* and *u* denote lower, medium, and upper numbers, respectively, to describe a fuzzy event. Additionally, the membership function *<sup>μ</sup>A* of a triangular fuzzy number can be expressed as follows:

$$\mu\_{\bar{A}}(\mathbf{x}) = \begin{cases} 0 & \mathbf{x} < l \\ \frac{(\mathbf{x} - l)}{(m - l)} & l \le \mathbf{x} \le m \\ \frac{(\mathbf{u} - \mathbf{x})}{(\mathbf{u} - \mathbf{u})} & m \le \mathbf{x} \le u \\ 0 & \mathbf{x} > u \end{cases} \tag{1}$$

where *l*, *m* and *u* are real numbers and *l* ≤ *m* ≤ *u*.

In view of above, the model of triangular fuzzy numbers is shown in Figure 2. The correspondence between the linguistic terms and triangular fuzzy numbers can be determined by Table 2. For any of two triangular fuzzy numbers <sup>∼</sup> *<sup>A</sup>* = (*l*1, *<sup>m</sup>*1, *<sup>u</sup>*1) and <sup>∼</sup> *B* = (*l*2, *m*2, *u*2), the operational laws of the two triangular numbers are as shown below:

$$\begin{cases} \tilde{A}\_1 + \tilde{A}\_2 = (l\_1 + l\_2, m\_1 + m\_2, \mu\_1 + \mu\_2) \\ \tilde{A}\_1 - \tilde{A}\_2 = (l\_1 - l\_2, m\_1 - m\_2, \mu\_1 - \mu\_2) \\ \tilde{A}\_1 \times \tilde{A}\_2 = (l\_1 \times l\_2, m\_1 \times m\_2, \mu\_1 \times \mu\_2) \\ \tilde{A}\_1 \div \tilde{A}\_2 = (l\_1 \div l\_2, m\_1 \div m\_2, \mu\_1 \div \mu\_2) \\ \lambda \tilde{A}\_1 = (\lambda l\_1, \lambda m\_1, \lambda \mu\_1), (k > 0) \\ \frac{\bar{A}}{\bar{\lambda}} = (\frac{l\_1}{\lambda}, \frac{m\_1}{\lambda}, \frac{\mu\_1}{\lambda}), (k > 0) \end{cases}$$

(2)

where *l*1, *m*<sup>1</sup> and *u*<sup>1</sup> are real numbers and *l*<sup>1</sup> ≤ *m*<sup>1</sup> ≤ *u*1.

**Figure 2.** Triangular fuzzy number.

**Table 2.** Correspondence between the linguistic terms and triangular fuzzy numbers.


This section is not mandatory but may be added if there are patents resulting from the work reported in this manuscript.

The main steps of the fuzzy DEMATEL method are briefly described as follows:

Step 1: Establish the fuzzy direct relation matrix *T* with fuzzy linguistic terms.

Step 2: Defuzzified-Initial relation matrix *F*. In this step, the fuzzy direct relation matrix *T* is defuzzified, namely the triangular fuzzy numbers are converted to crisp numbers by centroid method, a kind of defuzzification approach. Correspondingly, the Defuzzified-Initial relation matrix *F* can be established by Equation (3).

$$F\_{\vec{\mathcal{K}}}(\mathbf{x}) = \frac{\sum\_{i=1}^{n} \mathbf{x}\_{i} \mu\_{\vec{A}}(\mathbf{x}\_{i})}{\sum\_{i=1}^{n} \mu\_{\vec{A}}(\mathbf{x}\_{i})} \tag{3}$$

Step 3: Establish the normalized direct-relation matrix *X*. In this step, the initial direct-relation matrix *F* is normalized by utilizing Equations (4) and (5). Consequently, the normalized direct-relation matrix *X* can be obtained.

$$K = \frac{1}{\max\_{1 \le i \le n} \sum\_{j=1}^{n} a\_{ij}} \tag{4}$$

$$X = K \times F \tag{5}$$

Step 4: Establish the total relation matrix *M*. In this step, the total relation matrix *M* is calculated through Equation (6) where *I* denotes identity matrix. The element *mij* denotes the indirect effects that criterion *i* have on criterion *j*, and the matrix *M* gives the total relationship among the each pair of factors.

$$M = X(I - X)^{-1} \tag{6}$$

Step 5: Get the sum of rows and columns. In this step, the sum of rows and columns of matrix *M* are calculated through Equations (7) and (8). In the two equations, *ri* denotes all direct and indirect influence given by criterion *i* to all other factors, *cj* denotes the degree of influenced impact.

$$\tau\_i = \sum\_{1 \le j \le n} t\_{ij} \tag{7}$$

$$x\_j = \sum\_{1 \le i \le n} t\_{ij} \tag{8}$$

When *i* = *j*, (*ri* + *cj*) denotes all effects that are given and received by criterion *i*. (*ri* + *cj*) can show the degree of importance that criterion *i*, in the total system, namely the centrality of the element *i* in the problem group. Meanwhile, (*ri* − *cj*) represents the net effect that criterion *i* has on the system. If (*ri* − *cj*) > 0, the element *i* will be classified into cause group. By contrast, if (*ri* − *cj*) < 0, it will be classified into effect group.

Step 6: Establish the cause–effect relation diagram. In the final step, the cause and effect relationship diagram is depicted according to the dataset of (*ri* − *cj*). The horizontal axis (R + C) is obtained by adding R to C, and the vertical axis (R − C) is obtained by subtracting C from R.

Step 7: According to the results of the step 6, the cause group of the key drivers are ranked again to determine the most critical drivers. The calculation formula is:

$$R\_{\mathcal{G}} = \frac{(r\_i - c\_j)}{(r\_i - c\_j)\_{\text{max}}} \times 100\% \tag{9}$$

#### **5. Case Study: An Explanation**

*5.1. Case Background*

Due to the expanding trend of economic globalization, as well as the continuous development of smart and green technologies, CAMI must be able to respond to the market in a timely manner while taking into account economic growth, environmental protection, and the realization of social expectations. K Company is a medium-scale automobile manufacturing company in China, which has been committed to manufacturing energysaving and environmentally friendly vehicles. However, the automobile company found that the effect of implementing CER on energy conservation and emission reduction is far behind that of well-known foreign automobile manufacturing companies. Therefore, the company realized the need to evaluate the drivers of CER to ensure the effective implementation of its CER. The main research goals of the automobile company are as follows: (1) The automobile company hopes to understand the common drivers and key drivers of CER. (2) The company hopes to understand the comprehensive impact of implementing CER on economic, environmental and social benefits. (3) Last but not

least, automobile companies hope to achieve their own green development and sustainable development by effectively implementing CER throughout the entire life cycle of the car.

#### *5.2. Results and Analysis*

According to the results of the questionnaire, a fuzzy direct relationship matrix *T* in the form of fuzzy linguistic terms can be established. The fuzzy direct relationship matrix *T* is shown in Table 3. Using the centroid method, the fuzzy direct relationship matrix *T* is defuzzified into clear numbers, and transformed into the initial matrix *F* (Table 4), and the normalized direct relationship matrix *X* (Table 5) is established according to *F*, and then the total relationship matrix *M* is established (Table 6), and the row sum and column sum of the total relationship matrix *M* are calculated. Finally, a causality diagram is established, the results of (R + C) and (R − C) (Table 7) are calculated and used as the horizontal and vertical axes of the causality diagram, as shown in Figure 3.

**Table 3.** The fuzzy direct matrix T—the TBL's perspective.


**Table 4.** Initial relation matrix F—the TBL's perspective.



**Table 5.** The normalized direct-relation matrix *<sup>X</sup>*—the TBL's perspective (×10<sup>−</sup>2).

**Table 6.** The total relation matrix M—the TBL's perspective (×10<sup>−</sup>2).


**Table 7.** The values of R, C, (R + C), (R − C)—the TBL's perspective.


**Figure 3.** Cause–effect relation diagram.

The final results of our research are shown in Table 7 and Figure 3. According to Equations (7) and (8), we can know the elements in the cause group and the result group. The cause group includes drivers A1, A2, A3, A5, A10, A11, A12, A13, A16, and the effect group includes drivers A4, A6, A7, A8, A9, A14, A15. The results show that government regulations (A2), competitive advantage (A11), green supply chain pressure (A13) and green technology innovation (A5), incentive measures (A1) and standards (A3) are the six key drivers that promote the effective implementation of CER in the automotive industry. The most important driver is government regulations (A2). Half of the six key drivers are policy drivers. The fact is also true. The effective implementation of CER requires the government to formulate and supervise the implementation of environmental protection regulations and standards. Green supply chain pressure (A13), green technology innovation (A5) and competitive advantage (A11) rank second, third, and fourth among all important drivers. From the perspective of the effect group, it can be seen that media pressure (A15) ranks first in the effect group, and employee demand (A7) ranks last in the effect group.

As far as we know, there is no relevant analysis on the research directions involved in this paper, so it is impossible to give appropriate horizontal comparison results, but we have found similar experimental results in papermaking enterprises, industrial enterprises and the fashion industry. Among them, papermaking enterprises provide internal and external drivers for the company's green and sustainable development [56]. External drivers include government pressure, social pressure and economic pressure. Internal drivers include management, employees, corporate culture, the size of the company, and the financial situation. Experimental results show that economic pressure is the first driving force, and internal management and employee environmental awareness are the second driving force. The external factors for the green development of industrial enterprises [57] include policy and institutional environment, market environment and public supervision, and internal factors include the tangible and intangible resources of the enterprise. In the fashion industry [58], there are similar results. The driving factors of the sustainable fashion industry are attributed to internal driving factors (entrepreneurial direction and founder

culture, integration between different companies, innovation) and external driving factors (regulation, consumer awareness, competitiveness). Obviously, these have confirmed that government supervision, policies and regulations are the most important external driving factors for the manufacturing industry to fulfill its environmental responsibilities and implement environmental behaviors, which is basically consistent with the research results of this paper.

The verification is based on feedback from experts in the CAMI and references to relevant existing literature. After verification, our research results will be submitted to K Company.

#### **6. Discussion**

According to the research results, we drew a histogram of the key drivers and sorted the key drivers according to formula (9), as shown in Figure 4 and Table 8. It can be seen from Figure 4 and Table 8 that government regulations (A2) is the most important driver, ranking first. Only when the government and regulatory agencies jointly promote the automobile manufacturing industry to perform environmental responsibilities in strict accordance with regulations and standards can the company's environmental and economic interests achieve balanced development. In addition, green supply chain pressure (A13), green technology innovation (A5), and competitive advantage (A11) rank second, third, and fourth. These are the key drivers for the company to implement CER. These identified key drivers are interrelated: the implementation of government regulations (A2) is conducive to the improvement of enterprises' green technology innovation (A5), and promotes the improvement of enterprises' competitive advantage (A11), and the green supply chain pressure (A13) can also promote enterprises to comply with the laws and regulations of higher-level government departments, and improve their compliance and compliance. The other five drivers, including standards (A3), incentives (A1), consumers demand (A12), market trend (A16), and company image (A10), are also important for implementation of CER, and they all have different degrees of each other. The following research will conduct a correlation analysis of the importance of all these key drivers.

**Figure 4.** Causes group histogram.


**Table 8.** The sequencing of the key drivers.

In this study, we discussed the conclusions of the study and the importance of implementing CER, and drew some management implications including:

First of all, in order to accelerate the effective implementation of CER in CAMI, the government needs to improve the implementation and supervision of environmental protection policies and regulations, and increase the penalties for violations of environmental protection regulations. At the same time, it also encourages and praises companies that effectively implement CER and promote environmental improvements to promote its excellent practices. Second, the pressure of the upstream and downstream supply chains of automobile manufacturers also provides a strong impetus to promote the implementation of CER, because downstream consumers are more inclined to choose environmentally friendly products and are willing to pay for environmental protection. Upstream suppliers need to promote the design and development of more environmentally friendly and green products in the automobile manufacturing industry, thereby promoting the green innovation of upstream suppliers. Third, the green technological innovation of CAMI is also a very key driver, which is fundamentally the way for the enterprise to improve environmental performance and economic efficiency. Innovation is the source of all development; enterprises need to improve the construction of green technology innovation talent teams and increase the investment of green innovation costs. Fourth, the establishment of high-level environmental protection standards by the government and local governments is also a driving factor for enterprises to effectively implement CER. Finally, corporate incentives, public awareness of environmental protection, market trends, and corporate image can all have a positive impact on the implementation of CER by companies.

#### **7. Conclusions and Future Work**

Based on the perspective of TBL, and through the fuzzy DEMATEL method, we identified six key drivers from 16 common drivers, including government regulations (A2), green supply chain pressure (A13), green technological innovation (A5), competitive advantage (A11), standards (A3), and incentive measures (A1). These six key drivers are critical to the implementation of CER in CAMI. The main measures include the following: automobile manufacturers should improve the level of green technology innovation, pay attention to the needs of the upstream and downstream supply chain, improve the level of standard implementation, pay attention to competitive advantages, and formulate internal incentive mechanisms on the premise of meeting the requirements of government regulations. Only by continuing to meet the needs of all relevant parties can the company's environmental and economic benefits be comprehensively improved, and the company's green and sustainable development can be promoted. Through this research, K company can better understand the importance of CER practice to its own green development and sustainable development, and it is feasible to realize the coordinated development of economy, environment and society.

The work of this paper provides a valuable reference for the research and practice of CER in CAMI, finds the key drivers for the implementation of CER, and gives some management enlightenment. It is worth noting that this study still has certain limitations.

First of all, in the process of identifying and evaluating drivers, the number of questionnaires is relatively limited. Secondly, due to the limited amount of data, the proposed method cannot be further and extensively verified. The verified situation of a small and medium-sized automobile manufacturing industry cannot fully represent the situation of China's entire automobile manufacturing industry. The research conclusions cannot be widely applied to all automobile manufacturing industries, automobile sales companies, etc. These may constitute the basic elements of future research.

Therefore, the future research direction is from the perspective of the impact of the development of artificial intelligence and smart manufacturing technology on CAMI. The research perspective of CAMI's implementation of CER can also be shifted from the perspective of TBL to other perspectives, such as technology and multiple stakeholders. Regarding model construction and selection of multicriteria decision-making methods, the existing fuzzy decision-making can be extended to more advanced intelligent decisionmaking models and decision-making algorithms. In addition, it may be very interesting to study the relationship between CAMI's implementation of CER on corporate sustainable development and green development, and how to improve the company's image and increase profitability.

**Author Contributions:** M.Z. and S.L. received the research, collected data, performed the analyses and wrote the paper. H.Z., W.Y., Z.J. and Y.L. provided support and helpful suggestions in setting up and revising the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors are grateful for the research support from the National Natural Science Foundation of China (No. 51975432, 52075396).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

