PERFORMANCE ANALYSIS WITH FINITE TIME QUEUING IN CELLULAR SYSTEMS

Abstract

In a cell under heavy traffic, i.e., high call arrival rate and/or high average call dwell time within the cell, all the resources of the cell (i.e., channels) becomes blocked. And so all the new call attempts also get blocked, because of non-availability of channels. As the traffic increases the blocking probability of calls also increases. The situation is further complicated by retrials made by blocked calls. Our aim is to make an estimate of the blocking probability of calls with increasing traffic. But the resulting system is too complex to be solved using a mathematical formula. In this dissertation an attempt is made to estimate the blocking probability of calls with increasing traffic using simulation methods. An attempt is also made to design a network model of cell that improves the performance of the network by reducing the blocking probability. But experience have shown that FIFO queues gives poor performance as now every call has to wait for a long time to connect through. So in the model a small queue of size NQ with a maximum queuing time tQmax is maintained, if during the time tQ,,,1a the call is able to obtain a channel, it is connected through, otherwise after time tQ,ma, it is removed. from the queue. Hence, the queue, inspite of having minimum interference in the cell, can improve the throughput of the cell, and also reduce the blocking