A New Phase-Type Distribution to Break Kleinrock's Independence Assumption.
Yu Chen, Xuanhan Liang, Xiaofeng Tao · arXiv (Cornell University) · 2020
In this paper, we first introduce a new phase-type $PH(\mathbf{p},\mathbf{\lambda})$ distribution. The basic properties of the $PH(\mathbf{p},\mathbf{\lambda})$ distribution are investigated, including its Markov chain representation and the non-denseness property. Based on this new distribution, a model to predict message delay distribution in message-switched communication networks (MSCNs) is presented. Specifically, we consider that message lengths are kept unchanged when they traverse from node to node in a network, This scenario breaks Kleinrock's independence assumption but corresponds to actual scenarios. Finally, simulation shows that our model provides close and much better prediction results than that under Kleinrock's assumption.