研究具有两类失效模式的D策略M/G/1可修排队系统，其中第一类失效是服务台在服务顾客期间发生的失效，第二类失效是服务台在空闲期间发生的失效，且两类失效模式的失效率不同.使用全概率分解技术和利用拉普拉斯变换与母函数等工具，从任意初始状态出发，讨论了系统队长的瞬时分布和稳态分布，获得了系统稳态队长分布的递推表达式与稳态队长的随机分解结果.进一步，在建立费用模型的基础上，通过数值计算实例讨论了使得系统在长期单位时间内达到最小值的最优控制策略D*，并在同一组参数取值下与服务台不发生故障时的最优控制策略进行了比较.
This paper studies the M/G/1 repairable queueing system with D-policy and two types of failure modes. The first type of failure is the failure of the service desk during the service of a customer, the second type of failure is the failure of the service desk during the idle period, and the failure efficiency of the two types of failure modes is different. By using the total probability decomposition, L-transform, generating function and other tools, the transient distribution and equilibrium distribution of the system queue-length are discussed from any initial state. The recursive expression of the steady-state queue length distribution and the random decomposition results of the steady-state queue length are obtained. Furthermore, on the basis of establishing the cost model, the optimal control strategy D* is discussed, which makes the system reach the minimum value in a long-term unit time by numerical examples. The optimal control strategy under the same set of parameters is compared with the optimal control strategy when the service desk does not fail.
[1] Balachandran K R. Control policies for a single server system[J]. Management Science, 1973, 19(9):1013-1018.
[2] Balachandran K R, Tijims H C. On the D-policy for the M/G/1 queue[J]. Management Science, 1975, 21(9):1073-1076.
[3] Lee H W, Beak J W, Jeon J. Analysis of the MXX/G/1 queue under D-policy[J]. Stochastic Analysis and Applications, 2005, 23(4):785-808.
[4] Lee H W, Kim S A, Lee S W. Analysis and cost optimization of the M/G/1 queue under the D-policy and LCFS discipline[J]. Stochastic Analysis and Applications, 2008, 26(1):39-59.
[5] Agarwal R P, Dshalalow J H. New fluctuation analysis of D-policy bulk queues with multiple vacation[J]. Mathematical and Computer Modelling, 2005, 41(2-3):253-269.
[6] 魏瑛源, 唐应辉, 余玅妙. 基于Min(N,D)-策略的M/G/1 排队系统的队长分布及最优策略[J]. 系统科学与数学, 2015, 35(6):729-744.
[7] 魏瑛源,唐应辉, 余玅妙. 延迟Min(N,D)-策略的M/G/1 排队系统的队长分布与数值计算[J]. 运筹学学报, 2016, 20(2):23-37.
[8] 兰绍军, 唐应辉. 具有Bernoulli反馈和Min(N,D)-策略控制的Geo(λ1,λ2)/G/1离散时间可修排队的可靠性分析[J]. 系统科学与数学, 2016, 36(11):2070-2086.
[9] Lan S J, Tang Y H. Analysis of D-policy discrete-time Geo/G/1 queue with second J-optional service and unreliable server[J]. RAIRO-Operations Research, 2017, 51(1):101-122.
[10] Lan S J, Tang Y H. Analysis of a discrete-time Geo(λ1,λ2)/G/1 queue with N-policy and Dpolicy[J]. Journal of Applied Mathematics and Computing, 2017, 53(1):657-681.
[11] 曹晋华, 程侃. 服务台可修的M/G/1排队系统分析[J]. 应用数学学报, 1982, 5(2):113-127.
[12] 唐应辉. 服务台可修的M/G/1排队系统的进一步分析[J]. 系统工程理论与实践, 1996, 16(4):45-51.
[13] 唐应辉, 唐小我, 赵玮. 单重休假MX/G/1可修排队系统(I)-些排队指标[J]. 系统工程理论与实践, 2000, 20(1):211-215.
[14] 余玅妙, 唐应辉. 多级适应性延误休假MX/G(M/G)/1可修排队系统的可靠性指标[J]. 运筹学学报, 2008, 12(3):103-112.
[15] Wang K H, Kuo C C, Ke J C. Optimal control of the D-policy M/G/1 queueing system with server breakdowns[J]. American Journal of Applied Science, 2008, 5(5):565-573.
[16] 刘云颇, 唐应辉. 多重休假中以概率P进入的M/G/1可修排队系统[J]. 系统工程学报, 2011, 26(5):718-723.
[17] 唐应辉, 朱亚丽, 吴文青. 修理设备可更换的N-策略M/G/1 可修排队系统分析[J]. 系统工程理论与实践, 2014, 34(3):746-754.
[18] 高丽君, 唐应辉. 具有Min(N,D)-策略控制的M/G/1 可修排队系统及最优控制策略[J]. 数学物理学报, 2017, 37A(2):352-365.
[19] Tang Y H. A single-server M/G/1 queueing system subject to breakdowns:Some reliability and queueing problem[J]. Microelectronics and Reliability, 1997, 37(2):315-321.
[20] 唐应辉, 牟永聪, 余玅妙. 第二类故障期间以概率p进入的M/G/1 可修系统[J]. 系统工程学报, 2012, 27(4):559-567.
[21] 唐应辉, 朱亚丽. 具有温储备失效的M/G/1 可修排队系统[J]. 系统工程理论与实践,2014, 34(4):944-950.
[22] 唐应辉, 刘金银, 余玅妙. N-控制策略且温储备失效的M/G/1 可修排队[J]. 系统工程学报,2015, 30(6):852-864.
[23] 蔡晓丽, 唐应辉. 具有温储备失效特征和单重休假Min(N,V)-控制策略的M/G/1 可修排队系统[J]. 应用数学学报, 2017, 40(5):692-701.
[24] 唐应辉, 唐小我. 排队论——基础与分析技术[M]. 北京:科学出版社, 2006.
