Performance Bounds and Suboptimal Policies for Multi-Class Queue

Abstract

In this paper, we consider a general class of a queuing system with multiple job types and flexible service facility. We use a stochastic control policy to determine the performance loss in multi-class M/M/1 queue. The considered system is originally a Markov decision processes (MDP). The author showed how to compute performance bounds for the stochastic control policy of MDP with an average cost criteria. In practice, many authors used heuristic control policies due to some hardness in computing and running mathematically optimal policies. The authors found bounds on performance in order to an optimal policy where the goal of this job is to compute the difference of optimality and a specific policy. In other words, this study shows that, the optimal bounds of the average queue length for any non-idling policies can be found by a factor of service rates.

The response time analysis of queuing model in cloud computing environment

“…In summary, both of the numbers of cloud applications and providers have kept gradually increasing for a couple of years. As a result, performance managing and guaranteeing the Quality of Service (QoS) have been ones of the most important aspects of clouding computing [8].…”

“…In that work, the web applications are modeled as queues, and the virtual machines are modeled as service centers; they could be distinguished into two groups of priority classes with each class having its own arrival and service rates. In addition, the author in [8] showed how to compute performance bounds for the stochastic control policy of Markov decision processes with average cost criteria. In other words, he found bounds on performance with respect to an optimal policy.…”

Comparison of performance measures in non-Markovian queues

“…Para la evaluación de la efectividad del método propuesto, compararon resultados experimentales numéricos extensos del modelo de aproximación con simulación de eventos discretos en el software Arena. Madankan [22] tomó en consideración una clase general de un sistema de colas con múltiples tipos de trabajo y facilidad de servicio flexible. Además utilizó una política de control estocástico para determinar la pérdida de rendimiento en la cola M/M/1 de varias clases.…”