With the rapid expansion of technologies like the Internet of Things and Artificial Intelligence, the digital supply chain has become a core pillar of value creation. Despite its importance, Iranian industries remain in the early stages of digital transformation. This study employs Fuzzy Cognitive Mapping (FCM) supported by Active Hebbian Learning (AHL) and Nonlinear Hebbian Learning (NHL) algorithms to analyze causal relationships among barriers, requirements, and outcomes of supply chain digitalization. Data were collected through expert interviews and questionnaires. Findings reveal that while AHL and NHL differ in estimating causal intensity, their structural patterns remain consistent. Scenario-based analysis demonstrates that coordinated strategies encompassing managerial and technological initiatives significantly improve performance indicators. By proposing an integrated framework, this study supports the transition toward digital supply chains and provides actionable insights for policymakers. This research fills the gap in expert-based cognitive modeling for digital transformation in developing economies, offering a robust tool for strategic decision-making.
Introduction
Accordingly, this study seeks to answer the following research question: How do managers and analysts cognitively interpret the causal relationships among barriers, strategic actions, and outcomes of supply chain digitalization, and which actions can most effectively improve digital supply chain performance? The transformations brought about by the Fourth Industrial Revolution—driven by technologies such as the Internet of Things, Artificial Intelligence, Blockchain, and Big Data—have profoundly reshaped organizational processes. These technologies have enabled the emergence of intelligent, data-driven ecosystems where decisions are made with greater speed and precision. In such environments, organizations are increasingly compelled to move toward digital integration. Among all organizational domains, the supply chain has been significantly influenced, giving rise to the concept of the digital supply chain, which seeks to enhance agility and transparency in complex environments. Despite these advantages, industries in developing countries such as Iran remain in the early stages of digitalization. Challenges, including technological limitations, infrastructure deficiencies, and lack of digital culture have hindered progress. While previous research primarily focused on technological dimensions, limited attention has been paid to the subjective perceptions of managers. Strategic decisions are strongly influenced by managers’ cognitive perceptions of barriers and requirements. Analyzing these causal relationships can provide valuable insights for designing effective managerial strategies. FCM, as a soft modeling technique, enables the analysis of complex causal structures in uncertain environments. By integrating fuzzy logic with network structures, this approach facilitates modeling human perceptions. Despite the expanding literature, several research gaps remain. First, prior studies often examined barriers and outcomes separately, ignoring feedback-driven relationships. Second, limited research has explicitly modeled expert cognition in this field. Third, the use of advanced learning mechanisms like AHL and NHL for optimizing causal weights has received limited attention in supply chain digitalization research. To address these gaps, this study develops an integrated FCM model to analyze the causal relationships among digitalization components in Iranian industries. The main contributions include developing an integrated causal framework, modeling expert cognition, and comparing AHL and NHL algorithms to improve model robustness.
Methodology
This study employs a mixed-method approach to model digitalization components. In the qualitative phase, key concepts were extracted via content analysis of expert insights. In the quantitative phase, these concepts were modeled using FCM, and causal relationships were optimized through NHL algorithms implemented in Python. The FCMPy toolkit was used to simulate and refine the model, enabling parameter estimation and scenario-based analysis. Initial weight matrices were constructed from expert evaluations using fuzzy linguistic scales. Defuzzification was performed via the center of gravity method, and NHL was applied to enhance accuracy and stability. Intervention scenarios were then analyzed, leading to the final stabilization of the cognitive model structure.
Findings
Simulation results identified seven key influential factors with the highest impact on digital performance: connectivity, lack of R&D capabilities, scalability, technological inaccessibility, sustainability, insufficient employee competence, and supply chain resilience. In contrast, political instability showed the least influence on performance outcomes. Scenario-based analysis revealed that scenarios focusing on organizational agility, digital skills, and organizational security had the strongest impact on flexibility and competitiveness. Conversely, scenarios targeting only policy and infrastructure barriers showed weaker performance, highlighting that policy interventions alone are insufficient without integration with operational components. These findings offer a practical framework for prioritizing digital transformation initiatives, emphasizing the need to focus on high-impact enablers to achieve measurable improvements in key performance indicators.
Discussion and conclusion
This study utilized an FCM framework to model the relationships among digitalization enablers and outcomes. The application of the NHL algorithm proved effective in balancing accuracy and stability, making it a powerful tool for analyzing nonlinear systems. Theoretically, the findings demonstrate that FCM offers a robust alternative to linear models, enabling nuanced exploration of interdependencies. Practically, results highlight that successful transformation depends on high-impact enablers such as connectivity, R&D capabilities, and employee competence. Scenario analysis revealed that interventions targeting agility and decision-making processes were more effective than those focused solely on infrastructure, emphasizing the need for multidimensional strategies. Future research is encouraged to develop localized frameworks for implementing digital supply chains in the Iranian context. Special attention to managerial perceptions and cultural resistance will be essential for shaping effective policies. Finally, decision-support tools based on FCM and intelligent learning algorithms can assist managers in evaluating intervention scenarios and selecting optimal transformation pathways.