A Low Computation Cost Algorithm to Solve Cellular Systems with Retrials Accurately
This paper proposes an approximate methodology for solving Markov models that compete for limited resources and retry when access fails, like those arising in mobile cellular networks. We limit the number of levels that define the system by aggregating all levels beyond a given one in order to manage curse of dimensionality issue. The developed methodology allows to balance accuracy and computational cost. We determine the relative error of different typical performance parameters when using the approximate model as well as the computational savings. Results show that high accuracy and cost savings can be obtained by deploying the proposed methodology.
KeywordsApproximate Model Retrial Queueing Exact Model Customer Retrial Service Probability
Unable to display preview. Download preview PDF.