论文部分内容阅读
针对复杂应用环境中网络新媒体服务系统的特点,提出一种事件驱动的动态服务组合策略及其在线优化算法,在保证各类业务服务质量(QoS)的同时,提高系统资源的利用率.通过定义不同类型的事件,驱动服务组合的动态调整,实现对各类业务QoS的保障和对业务需求变化的感知.构建基于半Markov切换空间控制过程的系统分析模型,利用模型的动态结构特点,提出一种结合随机逼近和策略迭代的在线优化算法.该算法不依赖系统参数信息,对环境具有良好的自适应性.仿真实验结果验证了算法的有效性.
In view of the characteristics of network new media service system in complex application environment, an event-driven dynamic service composition strategy and its online optimization algorithm are proposed to improve the utilization of system resources while ensuring the quality of service (QoS) Define different types of events to drive the dynamic adjustment of service composition, realize the QoS guarantee of all kinds of services and perceive the change of business demands.A system analysis model based on semi-Markov switching space control process is constructed, and the dynamic structural characteristics of the model are proposed An online optimization algorithm based on stochastic approximation and strategy iteration, which has good adaptability to the environment, does not depend on system parameter information, and the simulation results verify the effectiveness of the algorithm.