[1]高文斌,王 睿,王田丰,等.基于深度强化学习的 QoS 感知 Web 服务组合[J].计算机技术与发展,2022,32(06):92-98.[doi:10. 3969 / j. issn. 1673-629X. 2022. 06. 016]
　GAO Wen-bin,WANG Rui,WANG Tian-feng,et al.QoS-aware Service Composition Based on Deep Reinforcement Learning[J].,2022,32(06):92-98.[doi:10. 3969 / j. issn. 1673-629X. 2022. 06. 016]
点击复制

基于深度强化学习的 QoS 感知 Web 服务组合()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年06期
页码:
92-98
栏目:
系统工程
出版日期:
2022-06-10

文章信息/Info

Title:
QoS-aware Service Composition Based on Deep Reinforcement Learning
文章编号:
1673-629X(2022)06-0092-07
作者:
高文斌; 王 睿; 王田丰; 祖家琛; 胡谷雨
陆军工程大学 指挥控制工程学院,江苏 南京 210007
Author(s):
GAO Wen-bin; WANG Rui; WANG Tian-feng; ZU Jia-chen; HU Gu-yu
School of Command and Control Engineering,Army Engineering University of PLA,Nanjing 210007,China
关键词:
服务组合; 服务质量; 深度强化学习; Web 服务; QoS 感知
Keywords:
service composition; quality of service; deep reinforcement learning; Web service; QoS-aware
分类号:
TP301
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 06. 016
摘要:
传统单体式软件架构由于耦合性高、扩展性差的原因,难以适应如今用户需求频繁变动的开发场景。 随着服务化理念的深入推广,利用独立的 Web 服务进行组合成为解决这一问题的可行方案。 如何利用功能不同、服务质量( Quality ofService,QoS)迥异的 Web 服务,构建出满足用户功能性需求及非功能性需求的组合服务成为服务计算领域的一个研究热点。 提出一种基于马尔可夫决策过程的服务组合模型,并设计了基于深度强化学习的求解算法。 应用深度网络提升模型表现,可有效解决大规模服务组合场景中现有服务组合算法寻优能力差的问题;进一步针对传统强化学习 Web 服务组合模型中奖励值估计不准确的问题,提出了一种基于卷积神经网络计算奖励值的方法,对服务历史 QoS 信息加以充分利用,并在公共数据集上做了实验验证。 实验结果表明,基于深度强化学习的服务组合 ADR-WSC 算法在大规模服务组合问题中输出的组合服务 QoS 更优,算法运行时间更短。
Abstract:
Traditional monolithic software architecture is difficult to adapt to today爷 s development scenarios where user requirementschange frequently due to the high coupling and poor scalability. with the further promotion of the concept of servitization,compositingWeb-services with independence to develop software has become a feasible solution to this problem. How to use Web - services withdifferent functions and different quality of service ( QoS) to build a composite service that meets the functional and non-functional needsof users has become a hot research topic in the field of service computing. A service composition model based on Markov decisionprocess is proposed with a solution algorithm based on deep reinforcement learning. The problem that existing service compositionalgorithms are difficult to derive the optimal solution for large-scale service combination scenarios can be effectively solved by using deepnetworks. Furthermore,to address the problem of inaccurate estimation of reward value in traditional reinforcement learning based Web-service composition models,a method based on convolutional network is proposed to calculate the reward value to make full use of theservice’s historical QoS information. Finally,a performance simulation is done on the public dataset,and the simulation results show thatthe deep reinforcement learning-based service combination “Adaptive Deep Reinforcement Learning-Web Service Composition” ( ADR-WSC) algorithm has higher efficiency in the large-scale service composition problem. It has better performance in terms of running timeand combined-QoS

相似文献/References:

[1]戴雪梅 姜浩.基于带权图规划算法的语义Web服务组合[J].计算机技术与发展,2010,(03):67.
　DAI Xue-mei,JIANG Hao.Semantic Web Services Composition Based on Weighted Planning Graph Algorithm[J].,2010,(06):67.
[2]杭志 徐德智 胡春华.移动电子商务中的服务组合研究[J].计算机技术与发展,2010,(04):187.
　HANG Zhi,XU De-zhi,HU Chun-hua.Research of Service Component in M - Commerce[J].,2010,(06):187.
[3]戚艳军 徐光辉 李强[].Erlang算法在呼叫中心的应用研究[J].计算机技术与发展,2010,(06):179.
　QI Yan-jun,XU Guang-hui,LI Qiang.Research of Erlang Algorithm in Call Center[J].,2010,(06):179.
[4]李磊 董健全.基于博弈论的P2P激励机制的研究与设计[J].计算机技术与发展,2009,(05):5.
　LI Lei,DONG Jian-quan.Research and Design of An Incentive Mechanism of P2P Based on Game Theory[J].,2009,(06):5.
[5]李莹峰 邓晓衡.DHT网络中基于测量的QoS监控系统[J].计算机技术与发展,2009,(05):188.
　LI Ying-feng,DENG Xiao-heng.Measurement- Based QoS Monitor System in DHT Network[J].,2009,(06):188.
[6]王建强 吴辰文 李晓军.基于NS-2的Adhoc网络QoS仿真与研究[J].计算机技术与发展,2009,(11):183.
　WANG Jian-qiang,WE Chen-wen,LI Xiao-jun.QoS Simulation and Research of Ad hoc Network[J].,2009,(06):183.
[7]严娜 黄映辉.基于模板和上下文的语义Web服务动态组合[J].计算机技术与发展,2009,(12):89.
　YAN Na,HUANG Ying-hui.Semantic Web Services Dynamic Composition Based on Templet and Context[J].,2009,(06):89.
[8]王冠东 李心科.一种评价网格服务质量的指标体系[J].计算机技术与发展,2008,(01):40.
　WANG Guan-dong,LI Xin-ke.An Index System for Evaluating Quality of Grid Service[J].,2008,(06):40.
[9]张雪芹 周莲英.无线网络的跨层设计[J].计算机技术与发展,2008,(02):88.
　ZHANG Xue-qin,ZHOU Lian-ying.Cross- Layer Design in Wireless Network[J].,2008,(06):88.
[10]汤萍萍 王红兵.基于强化学习的Web服务组合[J].计算机技术与发展,2008,(03):142.
　TANG Ping-ping,WANG Hong-bing.Web Service Composition Based on Reinforcement -Learning[J].,2008,(06):142.
[11]赵莉,李蜀瑜.基于DEC_POMDP的Web服务组合优化算法[J].计算机技术与发展,2014,24(03):74.
　ZHAO Li,LI Shu-yu.Web Service Composition Optimization Algorithm Based on DEC_POMDP[J].,2014,24(06):74.

更新日期/Last Update: 2022-06-10