RESEARCH OF TASK COMPLEXITY DECISION SYSTEM FOR MANUAL ASSEMBLY TASKS USING FUZZY COGNITIVE MAPS TO CONSTRUCT BAYESIAN NETWORKS

With the evolution towards intelligent manufacturing, Industry 4.0 has revolutionized and gradually improved product quality and efficiency. When applied to the assembly sector, Assembly 4.0 has also attracted widespread attention. Assembly 4.0 requires lean production, automated learning, and the ability to solve real-time design decisions through technology and convert the direct labor of operators into management information and decisions. Therefore, to better assess the complexity of manual assembly tasks, we believe it is necessary to build a multi-view and multi-dimensional decision-making system from the perspective of manual assembly. For this purpose, we propose in this study, a Bayesian network model based on fuzzy cognitive maps. Simultaneously, we use Genie3.0 to build the Bayesian network model. For optimum quantization parameter values as well as conditional probability directions' optimization between the parent node and the child node of the Bayesian network, we use the max-min hill-climbing (MMHC) algorithm and the Expectation Maximum (EM) algorithm. The influencing factors can be managed through the Bayesian network using this model, and the users can fully understand the changing rules among task complexity factors. This model can also provide theoretical and technical support for users, starting from the initial stage of product conception down to the completion of production.