Modeling and Parameter Optimization of Statistical Priority-Based Multiple Access Protocol  

作　　者：Siying Gao Ming Yang Hui Yu 

机构地区：[1]Dept. of Electronic Engineering,Shanghai Jiao Tong University

出　　处：《China Communications》2019年第9期45-61,共17页中国通信（英文版）

基　　金：supported by national fundamental research key project (No. JCKY2017203B082)

摘　　要：The Statistical Priority-based Multiple Access Protocol(SPMA)is the de facto standard for Tactical Target Network Technology(TTNT)and has also been implemented in ad hoc networks.In this paper,we present a non-preemptive M/M/1/K queuing model to analyze the performance of different priorities in SPMA in terms of average packet loss rate and delay.And based on this queuing model,we designed a percentile scoring system combined with Q-learning algorithm to optimize the protocol parameters.The simulation results show that our theoretical model is closely matched with the reality,and the proposed algorithm improves the efficiency and accuracy in finding the optimal parameter set of SPMA protocol.The Statistical Priority-based Multiple Access Protocol(SPMA) is the de facto standard for Tactical Target Network Technology(TTNT) and has also been implemented in ad hoc networks. In this paper, we present a non-preemptive M/M/1/K queuing model to analyze the performance of different priorities in SPMA in terms of average packet loss rate and delay. And based on this queuing model,we designed a percentile scoring system combined with Q-learning algorithm to optimize the protocol parameters. The simulation results show that our theoretical model is closely matched with the reality, and the proposed algorithm improves the efficiency and accuracy in finding the optimal parameter set of SPMA protocol.

关 键 词：SPMA QUEUING model Q-LEARNING percentile SCORING system 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]