Paper Title
Evolving Trends in Reliability Prediction Methodologies, Tools and Techniques over Years and Perspective Research Areas

Abstract
This paper intends to envision the historical trends and advancements in the field of product reliability prediction and intuitively helps to foresee future techniques, likely disruptions when apprised with the lenses of conventional approaches. The scope includes wider reliability industry practices and trends to analyze. The paper takes the mixed approach of bibliographic analysis with selective deep dives to specific relevant literature to deliver the intended objective. The study utilizes popular databases SCOPUS, Web of Science, Science direct and VOS viewer as mapping tool to extract additional trends and information related to the topic. There had been conventionally two major approaches at high level; probabilistic and deterministic which deploys inputs from opposite ends of product development lifecycle continuum. Deterministic approach builds through basic engineering understanding, working principals and takes more of transfer function approach to predict reliability life of products and can be initiated at early design stage. While approach does not accommodate practical variations and/or noises leading to failures, and it’s often questioned. While probabilistic approach on other hand accounts for said uncertainties as its build on product failures data during actual usage. The probabilistic approach purely being statistical struggles to connect system level failure to the lowest level complying with engineering theories. Reliability field has evolved in its approaches to predict failures and improve life of product in use. It started with approach of weakest link analysis, standard catalog-based life for electronics components, use of traditional empirical data-based approaches then to further POF (Physics-of-failure) based approaches with encouraging efficiencies when mechanisms of failure are known. Later there is lot of work done on mixed approaches to overcome some of the deficiencies in existing models and certainly shown promising use and improves practicalities. While with recent development of internet-of-things (IoT) enhancing ability to acquire big data along with significant advancement in advance analytics area with evolving techniques like ML and AI; leading world to apparently new avenues to address the product reliability. Are these advances real? Would those take reliability field to different level or on different track? Does these pose threat which needs to be managed with right guidance to keep the focus on physics of failures? Or would the focus likely to shift from reliability to availability. All these questions need an investigation and has hidden research opportunities which needs to be explored. The discussion in this paper does not answer all questions raised but certainly expect to bring clarity, shall provide wider understanding and perspective and attempts to identify areas as probable research gaps in reliability area. Keywords - Reliability, DFR, POF (Physics-of-failure), Reliability life predictions, Life data analysis, Reliability tools