融合网络环境下实时与自适应的服务选择机制研究

In integration network environments, due to the exponentially increasing number of services, the computation time of service selection sharply increase. At the same time, complex service selection process and the dynamic change of service environments lead to a serious shortage of service selection self-adaptability. All these bring new challenges to the service selection problem in integration network environments. To cope with these challenges, the study focuses on the real-time and self-adaptive service selection in integration network environments. Based on QoS-aware service filtering, service multi-dimensions QoS prediction, self-adaptive service selection process model, fast and efficient service selection algorithm, the study intends to build a service selection mechanism with real-time and self-adaptation guarantee, which will provide theoretical foundation and technical support for the future integration network service support platform construction. In terms of the research results, the study will seek a grant under ￥ 660000, and publish 10 or more SCI papers and 2 or more CCF A/B/C conference papers.

在融合网络环境下，指数性增长的服务数量经常导致服务选择的时间开销急剧上升；同时，复杂的服务选择流程和动态变化的服务运行环境，频繁导致服务选择的自适应性严重不足，这为融合网络环境下的服务选择问题带来了新的挑战。为此，本项目重点解决融合网络环境下服务选择的实时性和自适应性问题，拟通过QoS感知的服务过滤、服务的多维QoS预测、自适应的服务选择流程建模、快速高效的服务选择算法等研究，构建一种融合网络环境下具有实时性和自适应性保障的服务选择机制，为未来的融合网络服务支撑平台建设提供理论支撑和技术支持。

在融合网络环境下，指数性增长的服务数量导致服务选择(组合)的时间开销急剧上升，同时复杂的服务选择流程和动态变化的运行环境导致服务选择的自适应性严重不足，这为融合网络环境下的服务选择带来了新的挑战。本项目围绕融合网络环境下实时与自适应的服务选择机制开展深入研究，取得了以下成果：.1）提出了一种基于多维QoS预测的服务选择算法，相关研究成果发表在IEEE Transactions on Services Computing（中科院1区、CCF B类）、Journal of Parallel and Distributed Computing（CCF B类）等SCI期刊上。.2）提出了一种信誉度感知的服务自适应选择算法，相关研究成果发表在ACM Transactions on Multimedia Computing, Communications and Applications（CCF B类）期刊及IEEE ICWS（CCF B类）会议上。.3）提出了一种网络感知的快速服务选择算法，相关研究成果发表在IEEE Transactions on Cloud Computing（中科院1区）、IEEE Transactions on Services Computing、IEEE Transactions on Emerging Topics in Computing、Future Generation Computer Systems等SCI期刊及软件学报上。.4）提出了一种基于服务认知水平的用户偏好计算方法，相关研究成果发表在IEEE Transactions on Network and Service Management及IEEE ICWS（CCF B类会议）上。.5）提出一种高可靠的云服务供应算法，相关研究成果发表在IEEE Transactions on Services Computing期刊上。.基于上述成果，本项目实现了2个原型系统http://webcloudsim.org和http://qoecenter.com；培养博士/硕士研究生12人；发表SCI论文19篇，CCF B/C类会议论文4篇及2篇软件学报论文；授权2项国家发明专利；部分成果获得2018年度吴文俊人工智能学会科技进步一等奖。本项目均超额完成各项指标。