A methodology to determine the resilience of a system to obsolescence during the design phase.

Technologies, evolving at an ever faster pace, particularly in the electronics and IT sectors, lead inevitably systems to face the problems generated by obsolescence. Clearly, the manufacturer must be able to design systems so that an obsolescence problem has the least possible damaging or undesired consequences. The fact that a system remains as long as possible unaffected by changes imposed by obsolescence is called resilience. It, therefore, determines the ability of the system to continue providing its services while having at least one obsolete component, function or technology. The objective of this research work is to propose a modeling methodology and tools to ensure resilience to obsolescence; this resilience being directly defined by design choices. The methodology is based on the exploitation of modeled dependencies in the architecture of a system. The mapping of dependencies is done using different models obtained by applying a system engineering methodology, ARCADIA. The dependency model is then transformed into Bayesian networks to predict quantitatively not only the consequences of possible obsolescence issues in a given architecture but also to qualify the alternatives. This methodology, called Obsolescence Resilience By Construct, ORByC, is applied to the study of the rail signaling system. [ABSTRACT FROM AUTHOR]