Level: Intermediate
To achieve process improvement results using experimental design, first reduce the error in the target process, then plan and execute an experiment suited to the purpose, and analyze the obtained data to find the optimal solution. This is then applied to the process to confirm the improvement, and if the improvement is insufficient, the optimal solution is revised to ensure results. The above approach is not simply experimental design, but a comprehensive experimental design that brings together a series of approaches in a comprehensive manner. To master this, it is important to have practical experience in the PDCA cycle of plan, do, analyze, and improve, even if it is virtual, and traditionally teaching materials such as paper helicopters and table golf have been used. However, these have the disadvantage that it is not possible to learn efficiently alone in a short amount of time.
In recent years, a fly ball simulator has been developed as an add-in software for JMP that can flexibly generate process data including errors, making it possible for even one person to efficiently learn comprehensive experimental design. In this study, we have devised an educational program that links the fly ball simulator with the JMP experimental design platform and provides hands-on learning of the PDCA of comprehensive experimental design, assuming a custom design (optimal design). In particular, it is characterized by the use of multi-level repeated experiments to regress and correct the optimal solution while dealing with nonlinear responses.
