This study was conducted during January-March 2026 through expert consultations and structured interactions with professionals from the Maharashtra region of the Indian AFSCs sector. A systematic and structured methodology was adopted to identify, analyze and model the barriers affecting RL implementation in circular AFSCs. The methodological flow (Refer Fig 1) is aligned with established ISM procedures and follows a sequential approach starting from barrier identification to model development and classification using MICMAC.
The identified barriers are presented in Table 1. The development of the structural self interaction matrix (SSIM) is based on inputs collected from a panel of 21 experts drawn from academia, industry and AFSCs practice. The selection of experts was guided by their domain knowledge, professional experience and involvement in supply chain, logistics, Agri business and food processing sectors. The number of experts is considered adequate and consistent with prior ISM based studies, where panels typically range between 10 and 30 experts. A panel size of 21 ensures diversity of perspectives while maintaining manageability and consistency in responses. The profile of the experts is presented in Table 2.
The data collection was carried out in a structured manner. Initially, the list of identified barriers was shared with the experts. This was followed by individual interactions through online meetings and structured questionnaires to capture pairwise relationships among barriers. Experts were requested to indicate the direction of influence between each pair of barriers using standard ISM symbols. The iterative validation ensured reliability and consistency in the final responses. The consolidated expert inputs were then used to develop the SSIM, which forms the basis for further ISM analysis.
Development of SSIM
The resulting SSIM is presented in Table 3. The matrix reflects the consolidated expert judgement and represents the direct relationships among the barriers considered in this study. The symbol ‘V’ represents that one barrier influences another. The symbol ‘A’ represents the reverse relationship. The symbol ‘X’ indicates mutual influence, while ‘O’ denotes no relationship.
Development of initial reachability matrix (IRM) and final reachability matrix (FRM)
Following the development of the SSIM, the next step involves converting it into the IRM, where qualitative relationships are transformed into a binary format for further analysis.
The conversion is carried out using standard ISM rules. Based on these rules, the symbols in the SSIM are replaced with binary values. For instance, where a barrier i influences barrier j, the corresponding cell takes a value of one, while the reverse cell takes a value of zero. Similarly, appropriate binary representations are assigned for other types of relationships. This process results in a square matrix consisting of ones and zeros, representing direct relationships among the barriers. The IRM is presented in Table 4.
The IRM is further refined by incorporating the principle of transitivity to ensure that indirect relationships among barriers are also captured. If a barrier influences another barrier, which in turn influences a third barrier, then the first barrier is assumed to influence the third barrier indirectly. By applying transitivity, the FRM (Table 5) is obtained. This matrix reflects both direct and indirect relationships among the barriers.
Level partitioning and ISM model development
The driving power in FRM is then examined in decreasing order to understand the relative influence of barriers within the system. These values are further utilised in the level partitioning process, which forms the next stage of ISM analysis.
Level partitioning is performed to establish the hierarchical structure of barriers. For each barrier, the reachability set (Rb”) consists of the barrier itself along with those it can influence, while the antecedent set (Ab”) includes the barrier itself and those that influence it. The intersection set is obtained as the common elements between the reachability and antecedent sets. Barriers for which the reachability set and intersection set are identical are assigned to the top level of the hierarchy. Once identified, these barriers are removed from further iterations and the process is repeated until all levels are determined.
Based on the FRM and level partitioning results, the conical matrix (Table 6) is developed by rearranging the barriers according to their hierarchical levels. This matrix presents the driving power and dependence power of each barrier in a structured form and serves as the basis for developing the ISM digraph.
The ISM digraph (Fig 2) is developed based on the conical matrix to visually represent the interrelationships among the identified barriers. The nodes represent the barriers, while the directed links indicate the influence of one barrier over another. Barriers positioned at the lower levels exhibit higher driving power, indicating their strong influence on other interconnected barriers.
The final ISM model (Fig 3) is obtained after removing transitive links from the digraph, thereby retaining only the most meaningful direct relationships among the barriers. This results in a streamlined hierarchical structure that improves clarity and interpretability.
MICMAC analysis
MICMAC analysis is performed to classify the identified barriers based on their driving power and dependence power, using the values derived from the FRM. The analysis groups the barriers into four categories, namely autonomous, dependent, linkage and independent variables, as illustrated in Fig 4.
