出荷後の品質実績による損失関数実証の試み

抄録

An understanding of the ideas and mathematics of the loss function does not always provide a real sense of its significance. A major factor is the lack of evidence based on follow-up field studies. A survey and study were carried out to verify the loss function through post-shipment quality data. Instead of comparing the quality level calculated by the loss function with the actual monetary loss resulting from defects found after shipment, including defects found during further processing by the customer, different indices were compared. The quality level was replaced by the S/N ratio obtained through evaluation of product functionality, and the monetary loss was replaced by an index based on the number of customer complaints. A comparison of these substitute indices across model groups of switches indicated a strong correlation, demonstrating the validity of the loss function. To determine specific actions to be taken on the basis of the loss function, however, it will be necessary to calculate the total loss by adding the action cost to the monetary loss, instead of using a substitute index. Further study will be necessary.