摘要
经典排队模型M/M/n,M/G/1,GI/M/n,GI/G/1,网络排队系统以及从这些排队系统中发展起来的各种休假排队系统,都假定顾客输入的时间间隔为独立同分布的随机变量或构成一个马氏链。许多场合,特别是在通讯中,遇到许多排队现象,"顾客"的输入常常出现一些与经典模型大不一样的情况,用分形理论(具体说用一个混沌变换)去刻划才能吻合得比较好。作为这方面工作的尝试,一般情形下,用马尔可夫骨架过程理论求出了这类模型的队长的瞬时分布。
Classical queueing models M/M/n,M/G/1, GI/M/n, GI/G/1, queueing networks and various vacation queueing systems originated from the above-mentioned networks, all assume that time intervals of customers input are independent and identically distributed variables or they can form a Markov chain. However, in some environments, and communication practices in particular, the customer inputs are not the same as in the classical models, and the fractural theory can be used (i.e. using a chaos transformation ) to describe these inputs. This new research will initiate a new developing period for queueing theory. In the tentative research the transient distribution of such models using Markov skeleton process theory is obtained.
出处
《铁道科学与工程学报》
CAS
CSCD
北大核心
2004年第1期94-96,共3页
Journal of Railway Science and Engineering
