Cost characteristics of system service with input controlled by Markov chain

The paper presents the definitions and calculations related to the limit cost characteristics for a system consisting of N computers serving computational tasks. The tasks arrive at random intervals, having the same probability distribution. The calculation times are also random with the same probability distribution for separate computers. Markov chain models the process of allocation of computational tasks to computers. Total cost of service per unit time and an estimate for the total cost of waiting are presented. The characteristics obtained can be used to formulate and solve relevant optimization problems.