Towards intelligent control of complex manufacturing systems

A characteristic of complex manufacturing systems is the presence and complicating phenomenon of uncertainty in its various forms. Inability to accurately model and control these phenomena often leads to suboptimal and costly operation policies leading to noncompetitive production systems. Traditional operations research modeling and optimization techniques may not incorporate this uncertainty satisfactorily or allow for a computationally efficient and timely (online) solution. Fuzzy decision-making systems have been used to explicitly model uncertainty in management decisions that are inherently vague, uncertain, and imprecise due to the nature of human-related systems. Goals and constraints may only be vaguely defined and some of the data may be inherently subjective and imprecise. These techniques including several within the family known as intelligent systems and controls or the so called intelligent information systems (intellisystems) are now being applied to a bewildering array of problems. The paper synthesizes these problems, reviews some of the most important and proposes a new integrated fuzzy-neuro online control approach for a broad class of these complex manufacturing systems