Conflict resolution in the markov chains

Abstract

Parallel stochastic processes in some systems can be considered as simultaneous functioning of several Markovian chains with coinciding sets of states. The states are interpreted as undivided recourses, because at every moment no more than one chain can be in these states. The situation where several chains simultaneously require transition to the same state, forms the conflict. Conflict resolution consists in the choice of the chain, that realizes the required transition, while the other chains, which also have required to reach the same state, are in waiting mode, i.e. they have lost time. In the present work the new algorithm of such conflicts resolution is suggested; it can be used in caseswhere known method are inapplicable.