3. Obsolescence forecasting techniques

A prerequisite for any proactive obsolescence management approach is to predict when the system elements of interest will become obsolete. Such forecasts can only be probabilistic, and depend on numerous system parameters, both endogenous and exogenous. Anticipating obsolescence makes it possible to take rational decisions that minimize negative impacts. Against this backdrop, numerous studies have been developed to create models for forecasting obsolescence risk. Two types of methods can be distinguished: methods based on mathematical models and methods based on machine learning.