Poisson Traffic Processes in RGSMP and Generalized Networks
Xuan Cheng
Abstract
Xuan Cheng
Abstract
In this paper, we present conditions under which the traffic processes in a generalized semi-Markov process with reallocation (RGSMP) are Poison processes, and givean easy-to-use criterion for the quasi-reversibility of queues which can be described byRGSMP's. By applying these results to a multiple class generalized network with generalservice time distributions, we prove that its stationary state distribution is of a productform and the traffic processes which represent the customers exiting the network arePoisson processes.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In this paper, we present conditions under which the traffic processes in a generalized semi-Markov process with reallocation (RGSMP) are Poison processes, and givean easy-to-use criterion for the quasi-reversibility of queues which can be described byRGSMP's. By applying these results to a multiple class generalized network with generalservice time distributions, we prove that its stationary state distribution is of a productform and the traffic processes which represent the customers exiting the network arePoisson processes.
Key concepts: Queue, Poisson distribution, Markov process, Class (philosophy), Computer science, Traffic generation model, State (computer science), Mathematics