Data Modeling and ML Practice for Enabling Intelligent Digital Twins in Adaptive Production Planning and Control
Alessandro Chiurco, Mohaiad Elbasheer, Francesco Longo, Letizia Nicoletti, Vittorio Solina
Abstract
Technological advancements in AI, IoT, and Simulation push the frontier of the industry 4.0 to realize intelligent Digital twins (DT) of the industrial systems. This growing interest in manufacturing DTs inspires solutions for the flexibility, reliability, and resilience of production plans. However, the effort in building a clear guideline for dealing with data and algorithms is still its infancy. This paper combines multidisciplinary knowledge in Machine Learning (ML), and Production Planning & Control (PPC) to facilitate the integration of the ML algorithms into Production systems’ DTs. The paper proposes an architecture-based workflow to introduce ML and data practitioners into the creation of intelligent DTs for adaptive PPC. The framework in this study is explained with a simplified industrial case study that uses Neural Networks, k-Nearest Neighbor, and the Symbolic regression algorithms to justify the utility of the proposed framework.