A naive optimization method for multi-line systems with alternative machines

The scheduling of parallel machines and the optimization of multi-line systems are two hotspots in the field of complex manufacturing systems.When the two problems are considered simultaneously,the resulting problem is much more complex than either of them.Obtaining sufficient training data for conventional data-based optimization approaches is difficult because of the high diversity of system structures.Consequently,optimization of multi-line systems with alternative machines requires a simple mechanism and must be minimally dependent on historical data.To define a general multi-line system with alternative machines,this study introduces the capability vector and matrix and the distribution vector and matrix.A naive optimization method is proposed in accordance with classic feedback control theory,and its key approaches are introduced.When a reasonable target value is provided,the proposed method can realize closed-loop optimization to the selected objective performance.Case studies are performed on a real 5/6-inch semiconductor wafer manufacturing facility and a simulated multi-line system constructed on the basis of the MiniFAB model.Results show that the proposed method can effectively and efficiently optimize various objective performance.The method demonstrates a potential for utilization in multi-objective optimization.