Level: Intermediate
【Presenter】
Akira Hososhima, CEO, Torin Consulting
[Presentation overview]
Nowadays, with calls for work style reform, the importance of improving the efficiency of experiments and analysis is increasing. A good way to demonstrate the power of JMP's experimental efficiency is to show how experimental data can be replaced with a definitive screening design (DSD) or custom design to significantly reduce the number of experiments. The response data at that time are picked up from the existing experimental data. When I find experimental data divided into multiple tables, I put them together in one table, perform multivariate analysis, visualize it with a profile, and notice the pitfalls of the OFAT (One Factor at a Time) method. get. We show that there are multi-objective optimization and robust optimization methods based on accumulation processing, mean and variance, for the idea of analyzing repeated experimental data by averaging.
When design of experiments is used in development, the existence of interactions is often unpredictable, and interactions are not uncommon. DSD has no confounding between main effects and interactions (2FI) and no confounding between 2FIs, and the number of experiments can be as small as twice the number of factors. This is a big advantage. What I learned from actually using DSD, failures that occur when the number of main effects + number of interaction terms approaches the number of factors, solutions by extended planning, and practically important papers related to DSD obtained from the JMP community and ASQ Report what you think, etc.
[Profile of the presenter]
At Yamatake Honeywell (currently Azbil), he held positions such as General Manager of FA Development Department, Executive Director of R&D Headquarters, Director and General Manager of Quality Assurance Promotion Headquarters, Advisor to Azbil Kinmon, and established Tobayashi Consulting. Areas of specialization include yield improvement and quality improvement through production data analysis, general statistical problem solving such as market failure prediction, robust design, optimization design, and experimental planning, design review, root cause analysis (RCA), and human error prevention. On-site guidance such as prevention and process improvement. His publications include The Essence of Net Business, published by JUSE 2001 (co-authored) [Received the Telecom Social Science Award], Successful Strategies for Practical Venture Companies, Chuokeizai-sha 2011 (co-authored), Book on Easy-to-understand Problem Solving. Nikkan Kogyo Shimbun 2014 (single author). The main paper is "Proposal of a workshop that promotes the transmission and reception of awareness of near misses and discomfort on the production line" Quality Control Society 2016 [Received the 2016 Quality Technology Award]. Main lecture "Proposal of a mechanism to visualize organizational factors that induce work mistakes and promote improvement" (Discovery-Japan 2018)
"Solving Quality Problems with JMP - Defect Analysis and Reliability Prediction in the Manufacturing Industry" (Discovery-Japan 2019)
