Adjudicators for diverse redundant components

In fault-tolerant components realized using replication, several alternate results are produce d, from which a single correct result must be derived. This is done by a mechanism, that we call an adjudicator (a concept more general than a voter), which uses as its inputs the individual results produced by the replicated subcomponents comprising the redundant component. This paper gives a rigourous definition of the adjudication problem, summarizes the existing literature on the topic, and investigates the use of probabilistic knowledge about errors/faults in the subcomponents of a fault tolerant component to obtain good adjudication functions. The concept of an optimal adjudicator is introduced and discussed.