HYPERPARAMETER OPTIMIZATION OF BAYESIAN PRIOR DISTRIBUTION THROUGH EQUIPMENT DEVELOPMENT IN O&G INDUSTRY: A LITERATURE REVIEW

The creation of new technologies poses many technical obstacles in the Oil and Gas (O&G) industry, including disruptive developments. The process of estimating reliability in the design process is not trivial and includes several factors of uncertainty, especially in Research and Development (R&D). Thus, the Bayesian methodology has been widely used in these cases where a broad background of information is not available. The maximum entropy method (ME) is a technique used for estimating the parameters of a Bayesian prior probability distributions of failure events. Therefore, this paper aims to perform a systematic literature review to identify the main optimization methods used for solving ME in the Bayesian context and in the O&G industry. Among the gaps identified, it can be seen that the O&G context is a vast application area to be explored. Thus, opening a broad possibility of studies of optimization methods to be tested on this problematic.