Channel Switching Strategies for multistep Markovian Controllable Queuing Systems Problems

Alexander Mandel, Viktor Laptin
This paper deals with a controllable queuing system in which the number of switching service channels monitor and modify at control time points spaced apart by a fixed time step. At transition from step to step, the intensity of the simplest incoming flow changes in accordance with a Markov's chain. The system is in a stationary mode between the steps. A cost function is the minimization of the total average cost of the system over a multi-step planning period. The problem is to find a channel switching strategy. The parametric structure of an optimal strategy significantly simplifies its construction