Nowadays, guaranteeing the highest product variety in the shortest delivery time represents one of the main challenges for most of industries. The dynamic contexts where they have to compete push them to quickly readapt their processes, increasing the need for reactive decision-support tools to identify targeted actions to improve performance. Starting from the analysis of existing decision-support tools separately adopting simula-tion or data mining techniques, a framework that combines Association Rule Mining (ARM) and simulation has been developed to capitalize on the benefits brought by both techniques. On the one hand, ARM supports companies in identifying the main criticalities that slow down production processes, such as different causes of stoppage, giving a priority ranking of interventions. On the other hand, data-driven simulation is used to validate the ARM results and to conduct scenario analyses to compare the KPIs values resulting from different configu-rations of the production processes. Once the best-impacting mitigating actions have been implemented, the proposed framework can be iteratively used to define an updated set of intervention areas to enhance, promoting continuous improvement. This data-driven approach represents the key value of the framework, guaranteeing its easy-to-readapt and iteratively application. Theoretical contributions refer to the use of simulation with ARM not only to validate relations but to perform scenario analyses in an iterative way, as well as to the novelty appli-cation in a low-tech sector. From a practical point of view, a case study in the fashion industry demonstrates the usability and reliability of the proposed framework.
Data-driven decision support tool for production planning: a framework combining association rules and simulation / Fani, V.; Antomarioni, S.; Bandinelli, R.; Bevilacqua, M.. - In: COMPUTERS IN INDUSTRY. - ISSN 0166-3615. - 144:(2023), p. 103800. [10.1016/j.compind.2022.103800]