Improved Distribution Mean Method Based on Weighted Processing
Junwei Tian, Yongxuan Huang · Journal of Xi'an Technological University · 2006
Conventional Distribution Mean Method(DMM) used to analyze simulation output data of a DEDS system is prone to the effects of initial conditions or man-induced factors.Therefore,an improved DMM based on weighted processing is proposed.Different methods are used to classify observed data.Different weight is assigned to each class of the observed data according to its data distribution.The observed data are processed with a weighted processing method.A single queue service example is simulated and the results show that the proposed new algorithm reduces the errors caused by man-induced factors and initial conditions.The length of confidence interval was reduced by 8.5% when α=0.95.Thus,the scientificity and validity of the proposed algorithm is verified.