On the Parameter Selection of a Query Flow Control System Using the Probabilistic Pushing Out Mechanism

Abstract

A telematics device is modeled as a two-stream Markov one-channel priority queuing system of finite capacity equipped with a probabilistic pushing out mechanism. Its probability, $$0 \leqslant \alpha \leqslant 1$$ , of pushing out is the control parameter of the queuing system. It is experimentally and theoretically proved that on the plane of the load factors for high-priority and low-priority traffic, there are regions within which each of the loss probabilities depends on parameter α in a linear manner. The shape of these linearity regions is studied in detail for the case of absolute and relative priority, for each of which the limits of applicability of the linear loss law are constructed. The optimal value of the pushing out probability α is obtained according to the criterion of the minimal probability of losing high-priority claims under the condition of limiting the probability of losing low-priority claims. The described method was used to remotely control robotic devices under the conditions of the Kontur space experiment onboard the International Space Station.