Analysis of a finite buffer model with two servers and two nonpreemptive priority classes
Analysis of a finite buffer model with two servers and two nonpreemptive priority classes
复制标题
DOI：
10.1016/j.ejor.2007.09.021
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
W. Feng;Masataka Umemura
中科院分区：
文献类型：
--
作者：
W. Feng;Masataka Umemura
本文分析了一个具有两个服务器和两个非抢占优先级服务类的有限缓冲区排队模型。到达流是独立的泊松过程，两类服务时间均服从不同均值的指数分布。两个服务器中的一个专门为具有高优先级的一类保留，另一个服务器根据非抢占式优先级服务调度为两个类提供服务。对于该模型，我们描述其动态行为的四维连续时间马尔可夫过程。应用递归的方法，我们提出了明确的表示为这个马尔可夫过程的稳态分布。然后，我们计算了Laplace-Stieltjes变换和两类客户的实际等待时间的稳态分布。并给出了与其它模型的数值比较结果。
In this paper, we analyze a finite buffer queueing model with two servers and two nonpreemptive priority service classes. The arrival streams are independent Poisson processes, and the service times of the two classes are exponentially distributed with different means. One of the two servers is reserved exclusively for one class with high priority and the other server serves the two classes according to a nonpreemptive priority service schedule. For the model, we describe its dynamic behavior by a four-dimensional continuous-time Markov process. Applying recursive approaches we present the explicit representation for the steady-state distribution of this Markov process. Then, we calculate the Laplace–Stieltjes Transform and the steady-state distribution of the actual waiting times of two classes of customers. We also give some numerical comparison results with other queueing models.