Ai Based, Automatic Production System Decomposition

The paper introduces a new approach for automatic plant decomposition, based on artificial intelligence (AI) techniques. A novel automatic solution, based on a generalized feature selection technique and on artificial neural network (ANN) training, was developed by the author. The main goal is to explore connections among parameters of a given database, and based on the modelled dependency sets, validated groups of parameters can form individual parts of the analysed system. Applying this technique to a production database containing data typically inherited from the shop-floor level through process monitoring systems (or based on virtual simulation models) results in groups of connected production parameters. Consequently, it allows decomposing a running or simulated manufacturing system into smaller, individual and autonomous components. Therefore, the approach can provide the basis for production system decomposition and reconfiguration, too.