Fault-tolerant mapping algorithms onto hardware structure of a multiprocessor system

An approach to ensuring fault-tolerance of multiprocessor systems is presented. The approach is based on arranging computations in a system in such a way that in the presence of faults of some processing modules (PMs) the tasks (which have to be executed by the system) can be reassigned for execution between the remaining nonfaulty PMs, thereby reactivating the system. This reactivation may be performed with the system operation quality degraded within acceptable limits. The problem of developing the methods of task reassignment (TR) which meet system requirements for fault-tolerance and optimize the given parameters of the system operation quality are considered. This goal is achieved by means of fault-tolerant mapping algorithms (which have to be executed by the system) onto its hardware structure