Abstract
Major features of stochastic control of a computational process in multiprocessor systems are discussed with an example of the problem of settling the conflicts between concurrent processes. Concordant optimal decision rules for conflict settling are explicitly found under various constraints to minimize the number of conflicts on finite and infinite planning horizons. Multicriteria versions of stochastic control problems for parallel processes are stated and a sufficiently general vector criterial controlled Markov chains apparatus is suggested for their investigation. Under lack of apriori information on process characteristics, adaptive conflict settling algorithms are proposed which are expedient and may be conveniently used in the practical tabled algorithms class.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
