Abstract
Quality of a used-product is often highly uncertain, impacts pricing decisions, and is influenced by many factors like age and usage. However, analysis of usage and its uncertain nature is not well understood. In this paper, joint effects of age and usage on quality of used/end-of-life products are analysed. To conduct this research, Product State Transformation Diagram is proposed and used to analyse quality of used-products under uncertain age and usage conditions. Subsequently, governing multivariate stochastic partial differential equation is derived and paired with operational conditions to develop Quality Degradation Model (QDM). Then, QDM is numerically analysed and compared with common age-based model. Results indicate that age-based models overestimate quality. Using one-factor-at-time approach, sensitivity of QDM to its parameters and inputs is analysed. In a real-life case study, QDM is applied to evaluate prices of used-cars. To validate QDM and to generate insights towards its accuracy, a pool of popular statistical and machine learning models are trained and compared with QDM. Results show that performance of QDM is comparable to known models like SVM, LR, ANN-MLP for this application.