Research On Scheduling Strategies Of Queue Models Based On The QoS

Quality of Service is widely used in the field of communication system, service, manufacturing industry and many other practical applications, which plays a crucial role. It has become an important factor for measuring whether the service is effective and reliable or not. In recent years, with the development of communication network system, a higher Qo S is becoming the target of next generation communication network. In fact, there are many factors which can affect the Qo S, and the scheduling rules is one of the most important factors. In order to maximize the Qo S, this paper constructs three queueing network models, and studies the indicators of the Qo S with different customer rejection mechanisms under FIFO and LIFO scheduling rules respectively. The three models are promoted on the basis of previous literatures, and they are new models which full of valuable in research.Firstly, we introduce the multiple vacation policy based on the M/M/1 queueing model. We research the Qo S indicator of block and pushout customer rejection mechanisms under the scheduling rule of FIFO and LIFO respectively in limited buffer space. By making analysis of system model and using the theory of Markov and Laplace Stieltjes Transform method, we get the mean waiting time of each control schemes and numerical analysis of the result.Secondly, we research the Qo S indicator of BL and PO customer rejection mechanisms under the schduling rule of FIFO and LIFO respectively based on M/G/1 queueing network model in finite buffer space. By using embedded Markov chain theory, Laplace Stieltjes transform and total probability decomposition technique, we get the mean waiting time and throughput under different control schemes and make numerical analysis of them through a special case.Finally, we research the Qo S indicator of timeout customer rejection mechanism under the schduling rule of FIFO and LIFO respectively based on M/G/1 queueing network model in infinite buffer space. By using Level – crosing method, we get the mean waiting time under different control schemes and make numerical analysis of the results through a special case.