摘要
为了更具体地分析M/M/c和c个M/M/1并联系统在性能上的差异,首先分析了Little公式在应用中可能存在的缺陷,然后通过AnyLogic仿真工具对模型运行过程进行跟踪,最后通过管理系统仿真(general purpose simulation system,GPSS)JAVA仿真获取了2种排队系统中服务台利用率、平均队长、最大队长、平均等待时间等对比指标,并指出了M/M/1并联系统用解析法求解存在的缺陷.仿真结果表明:2种排队系统中服务台利用率几乎相同;M/M/c系统中顾客平均等待时间稍短于c个M/M/1并联系统,对传统排队论中的"与c个M/M/1并联系统相比,M/M/c系统可以显著提高服务效率和减少等待时间"结论进行了修正.此外,M/M/c系统中"短时等待"顾客更多,其"零等待"顾客数和"长时等待"顾客数均显著少于c个M/M/1并联系统.
To make a more detailed analysis of the differences in performance between M/M/c and M/M /1 c parallel systems,in this paper,first,the possible defects in the application of Little formula were analyzed. And then the running process of the model was tracked by Any Logic. Finally,the service desk utilization,average queue length,maximum queue length,average waiting time of the two queuing system by GPSSJAVA were obtained. The defects of the analytical method for the M/M /1 parallel system were identified. Simulation results show that the service desk utilization is almost identical to the two queuing systems. The average waiting time of customers in M/M/c system is slightly shorter than that of c parallel M/M /1 system. And the classic conclusion of 'compared with the c M/M /1 parallel system,M /M/c system can significantly improve the service efficiency and reduce the waiting time'conclusion was revised,Moreover,M/M/c system has more short term waiting customers,its zero waiting and long waiting customers are significantly less than that in c M/M /1 parallel system.
出处
《北京工业大学学报》
CAS
CSCD
北大核心
2016年第9期1324-1331,共8页
Journal of Beijing University of Technology
基金
北京市自然科学基金重点资助项目(8131001)
