Statistical Inference In Multi Server Queuing Model

With the development of electronic technology and information technology,people’s life has been inseparable from mobile phones and networks.However,the use of mobile phones cannot be separated from the service of operators.As a bridge between customers and operators,the business hall is particularly important.As one of the three major operators in China,China Mobile is a mobile communication operator with advanced user scale and network scale in the world,especially when 5g network is just emerging,there are more and more users and problems in service.For the business hall,it is not only to provide services for customers,but also to pay attention to customer experience and improve efficiency.It is very important to solve the problem of customer queuing.To solve this kind of problem,we must rely on the help of queuing theory.In this paper,Bayesian estimation is used to estimate the parameters of the queuing system of Jinan mobile business hall.The queuing model of the business hall is that the customer arrival time interval and service time are all subject to negative exponential distribution,and there are multiple servers.Bayesian estimation is based on the prior information to estimate the parameters.In the practical application of queuing theory,this method is found to be accurate in the queuing model,and the sampling method is simple,which greatly reduces the decision-making time of managers on the premise of ensuring the accuracy.First of all,we simulate the data of the queuing system and estimate the parameters,which proves that this method is feasible.Then we apply this method to the queuing model of the mobile business hall and get the estimated value of service intensity ρ in the model.According to this estimate,we calculated other performance indexes in the queuing model,and compared with the expected value of the manager,we found that the queuing model of the business hall met the manager’s expectation in the working day,but the queuing system in the non working day may have congestion,causing dissatisfaction of customers.According to the congestion situation,we add a counter on the basis of the original model.The simulation results show that this improvement will greatly reduce the possibility ofsystem congestion.In this paper,we find that it is feasible to make statistical inference by Bayesian estimation method,and it will reduce the research cost and time.Through the analysis of this paper,we put forward reasonable suggestions for the queuing system of a business hall in Jinan,and the managers can quickly adjust the number of counters according to the actual situation of the queuing system in the future.