A Study of Sustainable Product Design Evaluation Based on the Analytic Hierarchy Process and Deep Residual Networks

Abstract

Traditional product design evaluation processes are resource-intensive and time-consuming, resulting in unsustainably higher costs and longer lead times. Therefore, sustainable product design evaluation has become an increasingly crucial aspect of product design, focusing on creating a high-efficiency, high-reliability, and low-carbon-emission approach. In this study, we proposed an integrated approach that combines manual design evaluation based on the analytic hierarchy process (AHP) with an automatic design evaluation based on a ResNet-50 network in order to develop a sustainable design evaluation method. First, the evaluation level and indicators for the shape design of a tail-light were defined using the AHP. We followed this by establishing a determination matrix and weight coefficients for the design indicators to create a manual design evaluation model. Second, tail-light shape image datasets were manually annotated based on the evaluation indicators, and design datasets were constructed. The ResNet-50 algorithm was introduced to train the datasets, and the automatic evaluation model for product design was constructed through training and tuning. Finally, we validated the feasibility and effectiveness of the product design evaluation method, which was based on AHP and ResNet-50, by comparing the results obtained using both manual design and automatic design evaluations. The results showed that the proposed sustainable product design evaluation model provides an efficient and reliable method for evaluating product design, improves the decision-making process, and empowers the design and development process. The model enhances resource efficiency and economic sustainability.