This project aims to address the three bottleneck problems including the theoretical analysis problem of variable(K>=1) service polling system, the mathematical model analysis problem of the polling system with sleeping and the problem to obtain the accurate analytic solutions of the performance characters. By conducting in-depth investigation of the variable(K>=1) service polling system, this project will establish the mathematical model of the variable(K>=1) service system by employing the embedded Markov chain theory and the probability generating function. Then the effects of the variability of the K on the service model and the variation of the K when the system reaches steady status will be discussed in this project which lays the foundation for the further theoretical studying. This project will do deep analysis on the features of the polling system with sleeping to establish its mathematical model. And the one-order and high-order performance characters, such as the mean queue length, the mean cyclic period and the mean waiting time will be deduced. All the work can help us to explore the contribution of the sleeping for these performance characters above, and may light a way to find an effective method to obtain the accurate analytic solutions of the performance characters. We expect to achieve a breakthrough in current getting the approximate solution of these polling systems by simulation and provide the theoretical foundation for the applications and performance evaluation of polling systems. A lots experiments will be done to correct the method of sleeping factor estimation and optimize the range of K, so as to improve the theory model. With these findings, we plan to check its behaviour and adaptability by applying this service model to the design of the access control in data collection wireless communication networks. The conclusions drawn from this project will not only be useful for access control in communication networks, but also useful for polling scheduling in other control areas. This will extend the research and applications of the polling systems.
本项目拟解决轮询系统优化中的三个关键问题,即可变K(K>=1)服务轮询系统分析、可休眠轮询系统建模分析及特性量较精确解获取的有效方法问题。通过对可变K(K>=1)服务系统进行深入研究,采用概率母函数等方法,尝试进行建模分析,研究K值动态改变对系统模型产生的影响,探索稳定性能下K值的变化范围,为今后进一步的理论分析打下基础;深入分析可休眠轮询系统特征,建立服务模型,对平均排队队长、平均等待时间等一阶和高阶特性量进行解析,探讨休眠对特性量的贡献,寻求一种获得主要特性量较精确解的有效方法,突破目前对优化后的轮询系统采用仿真实验获取近似解的现状;最后通过实验改进休眠因子的参数估计方法、优化K值选取范围,完善理论分析模型,并在数据采集无线网络接入控制设计应用中检验其有效性和适应性。本项目的结论不但可用于通信网络接入控制设计中,也能用于其它相关控制领域的轮询调度,可拓展轮询系统的理论研究及实际应用。
本项目主要是对可休眠可变K轮询系统的特性、运行机制分析及其应用研究。.轮询控制策略常用于无线通信网络的接入控制、资源管理调度中,但由于系统复杂,在进行系统建模时,学者们往往忽略因为站点为空而触发的系统空闲状态以及由此触发的被动休眠状态,且对服务K值可变系统缺乏系统的分析。本项目结合休眠状态,采用嵌入Markov 链、概率母函数、状态方程等方法,首先建立了限定K(K=1)可休眠服务轮询系统的数学模型,并通过修正状态休眠函数对数学分析模型进行了优化,解决了在以往研究中由于忽略站点空闲状态或因空闲状态触发的休眠状态导致的数学模型中的状态缺失;获得了表征系统性能的一阶和高阶特性量,如系统的平均查询周期、平均等待时间等的精确数学解析,误差从近10%下降到几乎为0,从而验证了系统的空状态(或由此触发的休眠状态)对服务系统性能的影响。同时,获得了休眠因子对系统性能参数影响的闭式表达式,克服了已有研究普遍采用实验的方法或独立能耗模型来分析系统的局限性。其次,通过不断变化K值向门限、完全服务系统逼近,研究了可变K轮询系统的特性,获得了不同达到率下K的变化范围,为自适应调整可变K(K>=1)轮询控制策略的K值提供依据。此外,项目还对对两级优先级轮询系统的延迟特性和能耗特性进行了相关研究。在此基础上,将分析模型应用在WSN、WLAN、智能交通调度、智慧医疗等无线通信网络接入控制协议中,分析了休眠策略对系统性能的影响,验证了可休眠可变K(K>=1)策略的良好效果。相应成果发表在《Sustainability》、《Journal of Ambient Intelligence and Humanized Computing》、《Mathematical Problems in Engineering》、《仪器仪表学报》、《自动化学报》、《东南大学学报(自然科学版)》、《激光与红外》、《系统仿真学报》以及国际会议上。.本项目已取得的成果有22项,在国内外期刊上发表论文11篇,国际会议论文4篇,其中 SCE/EI 检索11篇,中文核心4篇;发明专利申请6项,其中授权2项;软件著作权1项。培养研究生已获工学硕士学位8人,在读硕士6人。
{{i.achievement_title}}
数据更新时间:2023-05-31
Protective effect of Schisandra chinensis lignans on hypoxia-induced PC12 cells and signal transduction
基于分形L系统的水稻根系建模方法研究
Intensive photocatalytic activity enhancement of Bi5O7I via coupling with band structure and content adjustable BiOBrxI1-x
Asymmetric Synthesis of (S)-14-Methyl-1-octadecene, the Sex Pheromone of the Peach Leafminer Moth
拥堵路网交通流均衡分配模型
轮询系统理论演进及应用控制协议研究
多接入点无线热点网络中优化选择接入控制研究
无线网络中轮询控制系统分析与改进的研究
面向组合服务的智能化接入控制关键技术研究