The textile and apparel industry, as one of the most polluting industries worldwide, has attracted considerable attention to sustainability-related issues. This industry, which is considered the second-largest polluter after the oil industry, has significant negative impacts on the three dimensions of sustainability, namely environmental, social, and economic sustainability. From an environmental perspective, activities in this industry contribute to air pollution, extensive greenhouse gas emissions, and serious damage to the environment and human life [1]. Sustainability challenges in this industry mainly arise from low environmental standards and inadequate social conditions for the workforce. This situation contrasts with the relatively high level of sustainability of resources and processes in industrial regions. Therefore, companies operating in the textile and apparel sector are striving to improve sustainability through effective measures, including reducing their carbon footprint and enhancing working conditions [2]. From an economic perspective, the textile industry is considered one of the largest industries worldwide after agriculture and contributes significantly to employment generation. The industry is of great importance to many countries, as it not only contributes financially to national economies but also provides numerous employment opportunities. At the same time, companies operating in the textile and apparel sector face requirements to reduce costs and project completion times, as well as challenges associated with geographical complexities [3]. Government policies and regulations also have a broad impact on the textile industry. For example, due to monetization policies, this sector has been affected to a greater extent than other sectors. In the present era, sustainable development has become an essential necessity [4]. Among the various challenges facing the textile industry, the assessment of the technologies employed is one of the key factors influencing the stagnation or development of the industry. The main focus of sustainable technology assessment is on the functions associated with the use of technology in the textile industry. This assessment requires strategies such as strategic and forward-looking planning in the textile industry, evaluating the economic feasibility of sustainable technology and the possibility of securing the necessary investment based on a resilient economy approach, and assessing the qualitative, functional, and technical considerations of technology in the textile industry. Accordingly, various countries around the world, particularly those with a long-standing presence in the textile industry, have shifted their production programs toward modern and advanced products. The continuation of production, updating of existing products, and development of new products are essential; therefore, various countries have incorporated these issues into their research agendas. In this context, risk management plays an important role in addressing the uncertainties and challenges associated with technology and sustainability. Risk management is a formal process that enables an organization to identify risks, analyze them, and respond to them in an appropriate and effective manner. In one of the most comprehensive definitions, the risk management process is defined as a systematic and logical process that includes the identification, analysis, and measurement of risks (collectively referred to as risk assessment), as well as addressing risk exposure by taking into account the context, or, in other words, the organization's internal and external environment [5]. Risk measurement and management, along with performance measurement, are essential for sustainable supply chain management, as violations of social and environmental sustainability practices pose risks to the brand image of the apparel and textile industry [6]. Therefore, identifying and evaluating the factors and risks affecting technology sustainability is of particular importance. Accordingly, the objective of the present study is to identify the factors and risks that have a significant impact on technology sustainability in the textile industry and to develop a comprehensive model for assessing and analyzing these factors through the use of Bayesian networks. The proposed model seeks to provide an appropriate tool for forecasting and assessing sustainable technology in the textile industry while taking into account existing uncertainties and complexities. The use of Bayesian networks, owing to their ability to analyze probabilistic relationships among various factors and evaluate different scenarios, constitutes one of the key aspects of this study and can provide valuable insights for improving technology management and supporting the sustainable development of the textile industry.

The present study is developmental in terms of its objective, as the proposed model contributes to the development of sustainable technology, the advancement of the textile industry, and technology assessment in the textile industry. In terms of data collection, the research method is descriptive, while in terms of the nature of the data, the study is qualitative [7], [8].

By examining the current situation, conducting library studies, and interviewing various experts, the most important strengths, weaknesses, opportunities, and threats affecting technology assessment in the country were identified. The identified factors were then presented to the experts. The items were designed based on a Likert-scale questionnaire, and the experts were asked to review the questionnaire and determine the importance of each factor in the field of technology assessment. Table 2 summarizes the interview questions.

Bayes’ theorem is the foundation of Bayesian inference and Bayesian networks. Bayes’ theorem was initially developed in the 1970s to update probabilities based on new information.

The following section examines Bayes’ theorem:

P(A|B)P(B) P(B|A) = , P(A)

where P(A) denotes the probability of occurrence of event A, P(B) denotes the probability of occurrence of event B, P(B|A) denotes the conditional probability of occurrence of event (B), given the occurrence of event (A), and P(A|B) denotes the conditional probability of occurrence of event (A), given the occurrence of event (B). For event (B), P(B) represents the prior probability and P(B|A) represents the posterior probability, each of which is associated with evidence. Therefore, Bayesian analysis describes the relationship between the prior and posterior probabilities using Bayes’ theorem.

A Bayesian network is a probabilistic graphical model that represents the relationships among a set of random variables and their conditional dependencies through a Directed Acyclic Graph (DAG). Bayesian networks consist of two components. The first component is the DAG, in which each node represents a random variable and the arcs represent conditional dependencies. For example, if there is an arc from node (A) to node (B), then (A) is the parent variable of (B). The second component is the conditional probability distribution associated with each node. If the variables are discrete, this distribution can be represented by a table that lists the probability that the child node takes each of its possible values for each combination of values of its parent nodes. It should be noted that strong conditional dependencies exist within a Bayesian network. Based on the chain rule, when the conditional dependencies are taken into account, the joint probability of all nodes in the graph can be expressed as follows:

P(C, S, R, W) = P(C)P(S|C)P(R|S, C)P(W|C, S, R).

Using conditional independence relationships, the above equation can be simplified as follows:

P(C, S, R, W) = P(C)P(S|C)P(R|C)P(W|S, R).

For a generalized model with (n) components X1, X2, … , Xi, … , Xn, when the parent node pa(Xi) of node Xi is given, (Xi) is conditionally independent of all other nodes except its descendants. Based on conditional independence, the joint probability can be expressed as follows:

P(X1 , X2 , … , Xn ) = ∏ P(Xi |pa(Xi )).

In this equation, (pa(Xi)) denotes the set of all parent nodes of (Xi).

However, if there is no conditional independence among the nodes, the joint probability is given by the following equation:

P(X1 , X2 , … , Xn ) = P(X1)P(X2|X1)P(X3|X1, X2) … P(Xn|X1, X2, … , Xn−1)

= ∏ P(Xi |X1, X2, … , Xn−1).

The number of independent parameters required in the above chain rule is 2n-1. When it is assumed that there are (m) parameters in the parent set of pa(Xi ), the maximum number of independent parameters required for the joint probability of a Bayesian network is 2m. In most normal systems, (m) is much smaller than (n); therefore, the computational complexity of Bayesian networks is significantly reduced.

The information required for modeling was obtained through an overall understanding of the textile industry, observations, a review of the literature, an examination of available analytical reports, interviews with managers and experts in the textile industry, and the opinions of experts at various technical and managerial levels. It should be noted that, among the factors introduced in the conceptual framework, some were incorporated into the model. In addition, different states were defined for nodes with multiple states based on the nature and characteristics of the variables. For example, “government programs and policies” was defined as a ranking-type node with three states: “efficient,” “acceptable,” and “inefficient.” Furthermore, for ranking-type nodes, a range from the lowest to the highest possible level was defined, with each level assigned a score ranging from the worst to the best condition. Table 6 presents the nodes used in the model.

To complete the Bayesian network, it is necessary to determine the relationships among the identified and defined variables. The relationships among the variables were examined through a review of the literature and previous research in the field of sustainability. In addition, the knowledge of experts in this field was used to verify and validate the identified relationships. Several face-to-face interviews were also conducted with experts in the textile industry. The AgenaRisk software was used for modeling. Fig. 1 presents the proposed

The implementation of the model using AgenaRisk software with the initial data is presented in Fig. 2.

In this section, two scenarios are considered to evaluate the model. The results obtained from the scenarios after implementing the model are reported below.

Sensitivity analysis can highlight the factors that significantly influence the output of the model. If the validation result is positive, it means that none of the possible causes for rejecting the validity of the model have been identified, and the validation procedure is considered complete. However, if the result of this process is not positive, it is necessary to conduct a verification process to identify the sources and reasons for the model’s lack of validity. For the research model, the results of the sensitivity analysis are reported as an example for the target variable.

The present study aimed to identify the factors and risks affecting technology sustainability in the textile industry and to develop a comprehensive model for assessing and analyzing these factors using Bayesian networks. By considering the uncertainty and complexities inherent in this industry, the proposed model provides an appropriate tool for forecasting and assessing sustainable technology. The use of Bayesian networks, owing to their ability to analyze probabilistic relationships among various factors and evaluate different scenarios, enables the provision of valuable insights for improving technology management in the textile industry. The results of this study showed that economic and social sustainability have a significant impact on technology sustainability in the textile industry. As economic sustainability increases, technology sustainability increases considerably.

Furthermore, improving social sustainability also has a positive effect on technology sustainability. The sensitivity analysis also indicated that certain factors have a substantial influence on the model output. These findings can help managers and policymakers in the textile industry focus on key factors and manage related risks to improve technology sustainability and ultimately contribute to the sustainable development of the industry.

The authors are grateful to all individuals who contributed to the completion of this study through their assistance and constructive input.

The authors declare that this study was conducted without financial support from any external organization.

The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.