A new risk analysis approach to seek best production action during new product introduction

Abstract

During new product introduction (NPI), to undermine losses and avoid potential risks, decision-makers oftentimes try out various alternative production actions until potential risks to manufacturing systems are satisfactorily reduced or eliminated. Very often, multiple process variables are involved in manufacturing systems that are the intuitive reflections of the performance of manufacturing systems. It is of great importance for decision-makers to extract information (e.g., the data concerning risk factors) from these process variables. To address the aforesaid concerns, a new risk analysis approach, grounding on the framework of three-way decisions (3WD) and failure modes and effects analysis (FMEA), is developed to seek the best production action. In the first place, the epidemic model is customized to characterize NPI, and then, the losses/severity (S) induced by alternative production actions can be attained. With the aid of the finite Gaussian mixture model (FGMM), we can judge whether a new failure mode occurs at arbitrary runtime. Thereafter, the detectability (D) and occurrence frequency (O) of new failure modes can be evaluated. A new dynamic Choquet integral is devised to seek the best production action by fusing S, D, and O. A case study, along with discussions, is provided to foreground the effectiveness and merits of the proposed approach.