Intelligent manufacturing for production planning aims at a faster production ramp-up. To advance this target, modified Hopfield neural networks for anomaly detection and classification are introduced. The model also enables recommendations to correct planning data, a similarity measure between production layouts, and interpretable decisions. The model is applied in process planning, synthetic data generation, and Leukemia diagnosis. Further development may achieve a fully autonomous planning.
Jan Michael Spoor
Künstliche Intelligenz HopfieldNetze Industrie 4.0 Wissensbasierte Systeme Produktionsplanung Artificial Intelligence Hopfield neural networks Intelligent manufacturing systems Knowledgebased systems Production planning