Simulation-based and ai-driven solutions for optimized management and maintenance of industrial systems

The scientific relevance underlying this doctoral dissertation focuses on the application of simulation methodologies and, more broadly, Artificial Intelligence–based algorithms, aimed at supporting industrial plant and equipment engineering and, more generally, the production domain as a whole. Aligned with these objectives, this dissertation covers themes ranging from the improvement of the most essential tasks within logistics processes, such as picking operations, supported by advanced simulation-based analyses, to the development of tailored algorithms capable of detecting failures and assessing wear levels in air treatment systems. It has been possible to identify useful insights. In particular, the adoption of advanced routing policies in picking operations enables energy reductions up to 36%, as well as the presence of linear relationships between the optimal operational leverages in fixed-time inventory management policies for different budget constraints. Similarly, the implementation of classification algorithms to support the adoption of predictive maintenance policies demonstrated anomaly detection performance of 85%. After an introductory Chapter, the Chapter 2 details the procedures employed and the methodologies developed. Chapter 3 reports the findings from the analyses performed, as well as the performance of the developed algorithms. In the Chapter 4 all the results will be discussed, and the work concludes with a final chapter reporting on international studies examining how these technological solutions i) are also impacting the field of education and ii) academic institutions can act as strategic partners for enterprises.