Forecasting near-future events in manufacturing cell control

Artificial neural networks (ANNs) are discussed with regard to their use in shop floor automation and materials flow control, especially as a part of a task-oriented intelligent manufacturing control system in which the manufacturing system is controlled by cooperative coexistence of its components. A prerequisite is local intelligence at each autonomous component. This means giving the workforce freedom and room to alter operation modes and to install equipment that can schedule near-future operations. The ANN is a nonlinear forecaster, and it requires on formal model. Once it is in place, the network continually learns from its environment and adapts to new patterns of data. These qualities make ANNs the superior choice for forecasting. It is demonstrated that the results of near-feature event forecasting by neutral networks with three hidden layers are better than those obtained by regression analysis and exponential smoothing