Consideration of uncertainties in a dynamic modeling system integrated with a deep learning based forecasting approach

Elements, including market demand, client requirements, and technological advancement of a product family in the current global competition, greatly influence the evolution of a product family. All of these impacting aspects must be completely comprehended for the dynamic modelling of a product family’s progression. However, the dynamic modelling of a complex system is often affected by (high) uncertainties due to the lack of information or measurement error. This study presents different sources of uncertainty in the area of product family. In this paper, the dynamic modelling system’s data uncertainty and parameter uncertainty are addressed by an advanced multi-criteria decision-making (MCDM) technique. An interval-valued fermatean fuzzy-based multi-attribute border approximation area comparison (IVFFN-MABAC) technique is used to pinpoint a product family’s key characteristics and support decision-making under uncertainty. This dynamic model is integrated with a deep learning-based forecasting model to predict the specifications of those critical properties of future products. For this application, Apple’s iPhone product family is taken into account as the case study. The numerical results validate the effectiveness of this approach. With the help of this method, the management will benefit from being able to identify the features that have a big impact on subsequent development and optimize the investment accordingly.