Abstract
In this paper, we propose to approximate the distribution and the moments of the number of events of a Markovian arrival process (MAP) in a time interval (0, t] by those of MAP events in an Erlang distributed random time interval. We present simple expressions for the later and show that when the order of Erlang distribution is high, the approximation is fairly accurate. The accuracy and computational efficiency of the proposed Erlangian approximation is compared with the uniformization method, and it is shown that the Erlangian method can be a useful tool in approximating the cumulative distribution function and moments of the number of MAP events.
Mathematics Subject Classification:
ACKNOWLEDGMENT
Support for Jiandong Ren from the Natural Sciences and Engineering Research Council of Canada (NSERC) is gratefully acknowledged.