The objectives of this project are investigation of chaotic or unpredictable behavior in manufacturing systems and development of methods for controlling variations in system performance due to nonlinear interactions. The novel approach identifies if a manufacturing systems is under control and prescribes control actions for systems not under control. Experimental evidence of chaotic or unpredictable behavior in manufacturing systems for certain scheduling rules is presented, and the transition in system behavior, from chaotic to periodic, due to change in scheduling disciplines is explained using theoretical results. Other system parameters influencing predictability of system performance are identified. Improvements in classical system performance measures, such as sojourn times, throughput rate, conformance to due dates, due to enhanced predictability are measured. Control schemes for manufacturing systems are presented that do not require explicit analytical models of the system. The research opens new directions for analyzing and controlling manufacturing systems, or in general any complicated system. The research is essential for developing an understanding of dynamic behavior of integrated manufacturing systems.
