Manufacturing engineering as an experimental science

Successful competition today requires design and production of high quality products with a broad range of features and on-time delivery. These market requirements lead to manufacturing requirements. The wide range of features leads to smaller lot sizes and on-time delivery leads to a requirement for reducing the cycle time variability. It is difficult to meet these requirements while reducing unit costs. Large lot manufacturing is well-understood. The same is not true for small lot manufacturing. Experimental observations show great variability in cycle time and theoretical studies suggest that it will be extremely difficult to develop control methods that will provide predictable cycle time in small lot manufacturing. These considerations have suggested that an experimental approach for developing control methods for small lot manufacturing may be more appropriate. This paper discusses our three phase approach for developing control methods for small lot manufacturing. This approach uses simulation to develop heuristic rules, emulation to try them out and real-time control in implementation. We believe that this will also provide a good means for providing students with experience in operating lines under realistic conditions and experience in how to implement these methods in practice.