一种自适应机制与决策策略的研究应用  

An Adaptive Mechanism and Decision-Making Strategy Research

作　　者：齐金平[1] 苏世雄 

机构地区：[1]兰州交通大学机电技术研究所,甘肃兰州730070

出　　处：《测控技术》2016年第3期118-122,共5页Measurement & Control Technology

基　　金："十二五"国家科技支撑计划项目(2012BAH20F05);甘肃省财政厅基本科研业务费(214153);兰州交通大学青年科学基金(2012012)

摘　　要：针对自适应系统中,何时触发系统调整以及如何选取最好方式满足系统,结合机器学习和效用理论的方法,提出一种自适应机制和基于复合优先级的决策策略,自适应触发机制是基于有限自动机和ECA规则相结合的基础之上改进而来,基于复合优先级的决策策略是结合软件质量属性和机器学习中效用理论的基础之上提出来的,最后将该机制和策略应用于在线视频点播系统(VoD)中,通过对该系统的分析,表明了自适应触发机制和决策策略的有效性。Considering that when to activate the system adjustment and how to select the best way to meet the system in self-adaptive system,an adaptive mechanism and a decision-making strategy based on the compound priority are proposed with the method of machine learning and utility theory.Self-adaptive trigger mechanism is improved on the basis of the combination of finite automata and ECA rules,and the decision-making strategy of compound priority is proposed by combining the software quality attributes and utility theory of machine learning.Finally,the mechanism and strategy are applied to online video-on-demand system（VoD）.The analysis of this system verifies the effectiveness of self-adaptive trigger mechanism and decision-making strategy.

关 键 词：自适应系统 机制 决策策略 效用理论 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]