摘要
对多服务台、缓存容量受限、用户等待太久后厌烦而中途离开的呼叫中心排队系统,运用马尔科夫决策过程理论,建立折扣准则最小期望代价函数模型,动态控制服务率。结果表明,合适的服务台数和到达率的情况时,最优策略会具有单调性,提高决策的效率,并且结合有限阶段和无限阶段进行分析,兼顾了系统稳态工作时服务台利用率。
In this paper, we present a queueing system model for inbound call centers which have multi-server and a finite number of places of buffer. The fraction of impatience customers who abandon are included. The objective is to control service rate dynamically and to minimize expecteddiscount cost functional equation over finite and infinite planning horizon by markov decision processes theory. Proof and numerical examples show that if the number of server and the service rate are suitable, the optimal decisions are monotonic on some assumption, which leads to an efficient optimization procedure. Furthermore, we consider the server running rate in a stationary state of the queueing system.
出处
《系统工程》
CSCD
北大核心
2007年第6期101-105,共5页
Systems Engineering
基金
国家高技术研究发展863计划资助项目(2005AA123910)
陕西省自然科学基金资助项目(2006F41)
