一种基于规则的软件体系结构层性能演化优化方法

发布时间：2018-03-16 07:03

本文选题：性能评估　切入点：性能优化　出处：《电子学报》2016年11期 　论文类型：期刊论文

【摘要】：目前基于规则的软件体系结构(Software Architecture,简记为SA)层性能优化方法大多未充分考虑优化过程中规则的使用次数和使用顺序的不确定性,导致了搜索空间受限而难以获取更优的性能改进方案.针对这一问题并以最小化系统响应时间为优化目标,文中首先定义一种基于规则的SA层性能优化模型RPOM,以将SA层性能优化抽象为求解最优规则序列的数学问题;然后设计一种支持SA层性能改进规则序列执行的框架RSEF;进一步提出一种采用约束检查、修复及统计学习机制的演化求解算法EA4PO;最后以Web应用为案例与已有方法进行实验对比.结果表明:(1)本文方法较已有方法可以获取更短的系统响应时间;(2)EA4PO所引入的统计学习机制可显著提高演化求解算法的收敛速度和解质量.
[Abstract]:At present, most of the rule-based software architecture (abbreviated as SA) layer performance optimization methods do not fully take into account the uncertainty of the number of rules used and the order of use during the optimization process. This results in limited search space and difficulty in obtaining better performance improvement schemes. To address this problem and to minimize system response time as the optimization goal, In this paper, a rule-based SA layer performance optimization model, RPOM, is first defined to abstract SA layer performance optimization into a mathematical problem for solving the optimal rule sequence. Then, a framework named RSEF is designed to support SA layer performance improvement rule sequence execution, and a constraint check is proposed. The evolutionary algorithm of repairing and statistical learning mechanism, EA4PO, is compared with the existing methods in the case of Web application. The results show that the proposed method can obtain a shorter response time than the existing methods and is introduced by EA4PO. The statistical learning mechanism can significantly improve the convergence rate and the quality of the evolutionary algorithm.
【作者单位】：福建师范大学软件学院;伦敦大学学院计算机科学学院;武汉纺织大学数学与计算机学院;湖南师范大学数学与计算机科学学院;
【基金】：国家自然科学基金(No.61305079,No.61370078,No.61402481) 武汉大学软件工程国家重点实验室开放基金(No.SKLSE 2014-10-02) 福建省自然科学基金(No.2015J01235) 福建省教育厅JK类项目(No.JK2015006) 湖南省教育厅资助科研项目(No.14C0680) 河北省自然科学基金(No.F2015403046)
【分类号】：TP311.5

【相似文献】