Methods to analysis of queueing models with state-dependent jump priorities

In this paper, exact and approximate approaches for studying queuing models with state-dependent jump priorities are developed. Both models with finite separate buffers and finite common buffer for heterogeneous calls are investigated. It is shown that both models might be described by twodimensional Markov Chains (2-D MC). Exact approach based on solution of appropriate system of balance equations (SBE) for state probabilities faced with big computational challenges for large scale models. To overcome the indicated difficulties an approximate approach based on the state space merging algorithms is developed. This approach allows to construct simple algorithms to calculate the Quality of Service (QoS) metrics of the examined models. The results of numerical experiments are demonstrated