Finding causal relationship and ranking of industry 4.0 implementation challenges: a fuzzy DEMATEL-ANP approach

Abstract

In recent years, the new technologies of the Industry 4.0 have been considered by the academic and industrial community due to the increasing sustainability and future competitiveness of the manufacturing sector. However, manufacturers face many challenges when implementing Industry 4.0 that must be identified and analyzed. The present study seeks to identify and analyze the challenges of implementing Industry 4.0 in the glass industry in Iran. Based on a systematic review of the literature, a total of 23 challenges were identified in 6 dimensions. In the next step, using the Fuzzy DEMATEL-ANP technique and a survey of 10 experts, causal relationships between challenges were identified and prioritized. Findings showed that financial challenges are the most critical cause, and organizational-managerial challenges are the most critical effect in implementing new technologies in Industry 4.0. Also, human-cultural challenges interact more with other challenges. Finally, the results showed the challenges of cyber security issues, lack of regulatory, legal and contractual mechanisms, lack of standards, laws and regulations and a general framework in the field of digital and lack of budget and financial instruments for education, R&D and digital operations, the most important challenges in the implementation of new technologies of the Fourth Industrial Revolution in polluting industries such as the glass industry. This study, for the first time, identifies and analyzes the challenges of implementing Industry 4.0 in one of Iran's polluting industries. The results of this study help policymakers to develop strategies for the acceptance and implementation of Industry 4.0.

Appendices

Appendix A

Fuzzy DEMATEL steps:

Step 1: Evaluation factors that are causal and usually involve many complex situations are formulated, and the fuzzy verbal scale is designed to deal with the ambiguity of human judgment, according to Table 2.

Step 2: Take the opinion of experts, and their average is calculated. To do this, by considering the number of expert P, the matrix P is obtained Z^p, …, Z², Z¹, each identified by the corresponding fuzzy numbers. To calculate the mean matrix, the relation $Z = \frac{{Z^{1} \oplus Z^{2} \oplus \ldots Z^{P} }}{P}$ is used. This matrix is called the "initial direct relation matrix"; Where ${Z}_{ij}=\left({I}_{ij},{m}_{ij},{u}_{ij}\right)$ (the value of each element of the Z matrix) are fuzzy triangular numbers. In addition, because the main diameter elements were zero, they are specified in the matrix as (0, 0, 0) (Table

Table 7 The fuzzy linguistic scale

7).

Step 3: Through Equation A.1, the standardization relation, the index scales are converted to comparable scales. In the following relation, the matrix X is called the "standardized direct relation fuzzy matrix":

$$ \begin{gathered} a_{ij} = \left( {\mathop \sum \limits_{j = 1}^{i} I_{ij} ,\mathop \sum \limits_{j = 1}^{i} m_{ij} ,\mathop \sum \limits_{j = 1}^{i} u_{ij} } \right) \hfill \\ r = \max_{1^\circ i^\circ n} \left( {\mathop \sum \limits_{j = 1}^{i} u_{ij} } \right) \hfill \\ x_{ij} = \frac{{Z_{ij} }}{r} = \left( {I_{ij}{\prime} ,m_{ij}{\prime} ,u_{ij}{\prime} } \right) \hfill \\ \end{gathered} $$

(A.1)

Step 4: The fuzzy total relation matrix (T) is obtained. Where ${x}_{ij}=\left({I}_{ij}{\prime},{m}_{ij}{\prime},{u}_{ij}{\prime}\right)$ And the values of the matrix elements ${X}_{u},{X}_{m},{X}_{I}$ Contain the values ${I}{\prime}$, ${m}{\prime}$ and ${u}{\prime}$ in the X matrix, respectively.

$$ XI = \left[ {I_{ij}^{\prime \prime } } \right],Xm = \left[ {m_{ij}^{\prime \prime } } \right],Xu = \left[ {u_{ij}^{\prime \prime } } \right] $$

(A.2)

Given that $t_{ij} = \left( {I_{ij}^{\prime \prime } ,m_{ij}^{\prime \prime } ,u_{ij}^{\prime \prime } } \right)$, we have:

$$ \begin{gathered} \left[ {I_{ij}^{\prime \prime } } \right] = X_{I} \times \left( {I - X_{I} } \right)^{ - 1} \hfill \\ \left[ {m_{ij}^{\prime \prime } } \right] = X_{m} \times \left( {I - X_{m} } \right)^{ - 1} \hfill \\ \left[ {u_{ij}^{\prime \prime } } \right] = X_{u} \times \left( {I - X_{u} } \right)^{ - 1} \hfill \\ \end{gathered} $$

(A.3)

In this relation, I is a unit matrix, X_I, X_m and X_u are each n × n matrix, the components of which form the lower number, the middle number and the upper number of the fuzzy triangular numbers X matrix, respectively.

Step 5. Determining row (Di) and column (Rj) sums In order to calculate the sums of row (Di), and column (Rj) for each row i, and column j in the total-relation matrix (T), that is obtained using the following equations.

$$\widetilde{\mathrm{D}}={\left({\widetilde{\mathrm{D}}}_{\mathrm{i}}\right)}_{\mathrm{n}\times 1}={\left[\sum_{\mathrm{j}=1}^{\mathrm{n}}{\widetilde{\mathrm{T}}}_{\mathrm{ij}}\right]}_{\mathrm{n}\times 1}$$

(A.4)

$$\widetilde{\mathrm{R}}={\left({\widetilde{\mathrm{R}}}_{\mathrm{i}}\right)}_{1\times \mathrm{n}}={\left[\sum_{\mathrm{i}=1}^{\mathrm{n}}{\widetilde{\mathrm{T}}}_{\mathrm{ij}}\right]}_{1\times \mathrm{n}}$$

(A.5)

Step 6. Determining the prominence and relation values. In this step, a cause-effect diagram is drawn. This diagram consists of two axes, including the horizontal axis (${\widetilde{D}}_{i}+{\widetilde{R}}_{i}$), which is drawn by adding R to D, and the vertical axis (${\widetilde{D}}_{i}-{\widetilde{R}}_{i}$), which is drawn by subtracting ${\widetilde{R}}_{i}$ from ${\widetilde{D}}_{i}$. The horizontal axis (${\widetilde{D}}_{i}+{\widetilde{R}}_{i}$) indicates the prominence based on the importance degree of sub-criteria used in the study.

On the other hand, the vertical axis (${\widetilde{D}}_{i}-{\widetilde{R}}_{i}$) represents the relation based on influence. If the relation is negative, the sub-criteria will be classified as the effect group, influenced by other criteria. Nonetheless, if the relationship is positive, the sub-criteria will be classified as the cause group, which has a significant impact on other sub-criteria.

Step 7. Defuzzification ${\widetilde{(D}}_{i}+{\widetilde{R}}_{i})$ and $({\widetilde{D}}_{i}-{\widetilde{R}}_{i})$ values obtained from the previous step, which is done according to the following equation. In this relation, the Defuzzification B is the number $\widetilde{\mathrm{A}}= \left({a}_{1},{a}_{2},{a}_{3}\right)$.

$$B=\frac{l+m+n}{3}$$

(A.6)

Appendix B

Fuzzy DANP steps:

Step 1: Create a network structure between the criteria and calculate the total-relation matrix between the criteria and sub-criteria using Fuzzy DEMATEL.

Step 2: Form a non-weighted Super matrix from the complete communication matrix: This step is done once for the total-relation matrix between criteria and again on the total-relation matrix between sub-criteria. If we call the total-relation matrix obtained by the Fuzzy DEMATEL method T, to obtain the unweighted supermatrix (${T}_{c}^{\alpha }$) we must first normalize the data in each block related to the sub-criteria of a criterion and enter the final matrix obtained. Putting blocks together is considered as a supermatrix weighted.

$$T=\left[\begin{array}{cc}{T}^{11}& \begin{array}{ccc}\cdots & {T}^{1j}& \begin{array}{cc}\cdots & {T}^{1m}\end{array}\end{array}\\ \begin{array}{c}\vdots \\ \begin{array}{c}{T}^{j1}\\ \begin{array}{c}\vdots \\ {T}^{m1}\end{array}\end{array}\end{array}& \begin{array}{c}\begin{array}{ccc} & \vdots & \begin{array}{cc} & \vdots \end{array}\end{array}\\ \begin{array}{ccc}\cdots & {T}^{ij}& \begin{array}{cc}\cdots & {T}^{im}\end{array}\end{array}\\ \begin{array}{c}\begin{array}{ccc} & \vdots & \begin{array}{cc} & \vdots \end{array}\end{array}\\ \begin{array}{ccc}\cdots & {T}^{mj}& \begin{array}{cc}\cdots & {T}^{mm}\end{array}\end{array}\end{array}\end{array}\end{array}\right],$$

(B.1)

$${T}^{11}=\left[\begin{array}{cc}{T}_{11}^{11}& \begin{array}{ccc}\cdots & {T}_{12}^{11}& \begin{array}{cc}\cdots & {T}_{1{m}_{2}}^{11}\end{array}\end{array}\\ \begin{array}{c}\vdots \\ \begin{array}{c}{T}_{21}^{11}\\ \begin{array}{c}\vdots \\ {T}_{{m}_{1}1}^{11}\end{array}\end{array}\end{array}& \begin{array}{c}\begin{array}{ccc} & \vdots & \begin{array}{cc} & \vdots \end{array}\end{array}\\ \begin{array}{ccc}\cdots & {T}_{22}^{11}& \begin{array}{cc}\cdots & {T}_{2{m}_{2}}^{11}\end{array}\end{array}\\ \begin{array}{c}\begin{array}{ccc} & \vdots & \begin{array}{cc} \ddots & \vdots \end{array}\end{array}\\ \begin{array}{ccc}\cdots & {T}_{{m}_{1}2}^{11}& \begin{array}{cc}\cdots & {T}_{{m}_{1}{m}_{2}}^{11}\end{array}\end{array}\end{array}\end{array}\end{array}\right],$$

(B.2)

By dividing the elements of each row by the matrix ${T}^{ij}$ by the sum of the elements of the same row and putting them together, the normalized matrix ${T}^{\alpha }$ is obtained. The unweighted supermatrix will be the transatrix of the matrix obtained in this step:

$${T}_{c}^{\alpha }={({T}^{\alpha })}{\prime}$$

(B.3)

Step 3: Formation of a supermatrix weighted: In this step, using the unweighted supermatrix, between the main criteria of the element related to each block, the supermatrix related to the sub-criteria is multiplied by the block elements, and the supermatrix weighted (W) is formed.

Step 4: Determining the absolute priorities: After forming the supermatrix weighted using the infinite limit of the formed matrix, the final weights are determined.

$$\underset{k\to \infty }{\mathrm{lim}}{W}^{k}$$

(B.4)

Appendix C

See Tables

Table 8 Fuzzy T matrix

8,

Table 9 Fuzzy normal matrix (intensity of direct relations) for dimensions

9.