市場データを用いた車両設計指標の構築と最適化設計に関する研究

抄録

In this research, we proposed a process in which criteria was built by Data Envelopment Analysis with vehicle data on the market. Using DEA, we derived a one-dimensional evaluation value from a multi-dimensional index. However, DEA has flexibility in evaluation model. So we constructed it with Interpretive Structural Modeling. ISM is the method to clarify relationship between items with paired comparison and create a hierarchical structure chart. As a result, we made 32 evaluation models and determine the best model which has the maximum standard deviation. This enabled us to eliminate the intention of the user. In addition, we obtained some features of the evaluation result. For example, it was difficult for vehicle data which has better fuel efficiency to acquire a high evaluation. Larger or smaller displacement vehicle obtained higher evaluation value, and middle class vehicle obtained middle value widely. In addition, we optimized the value of each evaluation item and obtained proposition for design, which was derived from data of 108 new vehicles and proposed larger or smaller displacement and shape. From the above, we confirm that DEA is useful for evaluating products and ISM is efficiency for constructing the evaluation model of DEA.