事件驱动的无线CPS实时消息调度的并行优化研究

发布时间：2018-06-22 08:30

本文选题：实时服务质量 + 并行优化　；参考：《南昌航空大学》2017年硕士论文

【摘要】：信息物理融合系统是一个综合计算、网络和物理现实的多维复杂系统,在交通、国防、能源、医疗、大型建筑等方面有着巨大的应用市场,实时服务质量是制约其发展的关键因素。在信息物理融合系统中,消息调度机制和策略决定了其实时服务质量的优劣,而消息调度的并行优化是设计消息调度机制和策略过程中的重要环节。合理的消息调度并行优化方法不但可以避免或减少信道竞争和碰撞冲突,还能增加消息调度过程的稳定性和可靠性,提高资源利用率,延长系统寿命。本课题以国家自然科学基金为来源,主要对信息物理融合系统中实时消息调度的并行优化进行研究。论文分析了无线网络事件消息的特点和影响无线网络信道利用率的因素,对现有无线网络消息调度方法进行了分类,归纳起来主要为基于竞争的方法、基于时分复用的方法、混合的方法、基于优先级的方法等,通过深入分析可知基于时分复用的消息调度方法更加符合信息物理融合系统应用的特点。论文从消息之间的并行性角度出发,结合时分复用的理论,提出了一种事件驱动的实时消息并行调度算法ERMPSA(Event-driven Real-time Message Parallel Scheduling Algorithm)。算法主要从并行消息的判定、消息传输的路径选择和消息的并行传输三个方面展开,将满足消息截止期限要求作为网络的首要目标,同时还考虑了节点剩余能量和网络平均延时。为了让尽可能多的消息并行传输,依据消息调度的并行优化模型为同一时隙内所需传输的消息建立相应的消息图,利用图着色理论为消息图进行多轮条件着色,使所用颜色数最少,相同颜色的顶点所对应的消息可以并行传输。实验结果表明,依据消息调度的并行优化模型所设计的ERMPSA算法能够最大化并行程度。和现有消息调度方法相比,ERMPSA算法在不增加节点能耗的情况下,能够有效降低消息截止期限失去率,在降低消息平均传输延时和提高网络吞吐量方面也有较好的表现。
[Abstract]:The information physics fusion system is a multi-dimensional complex system of comprehensive calculation, network and physical reality. It has a huge application market in transportation, national defense, energy, medical treatment, large buildings and so on. Real-time service quality is the key factor restricting its development. In the information physical fusion system, the message scheduling mechanism and policy determine the quality of service in real-time, and the parallel optimization of message scheduling is an important part in the process of designing message scheduling mechanism and policy. Reasonable parallel optimization method of message scheduling can not only avoid or reduce channel competition and collision conflict, but also increase the stability and reliability of message scheduling process, improve resource utilization and prolong system life. Based on the National Natural Science Foundation of China, the parallel optimization of real-time message scheduling in information physics fusion system is studied in this paper. This paper analyzes the characteristics of wireless network event messages and the factors that affect the channel utilization of wireless networks, classifies the existing wireless network message scheduling methods, which are mainly competition-based methods and time-division multiplexing methods. By analyzing the hybrid method and the priority-based method, we know that the message scheduling method based on time division multiplexing is more suitable for the application of information physics fusion system. Based on the theory of time division multiplexing and parallelism between messages, this paper proposes an event-driven real-time message parallel scheduling Algorithm). (ERMPSA). The algorithm is mainly developed from three aspects: the judgment of parallel messages, the path selection of message transmission and the parallel transmission of messages. The primary goal of the network is to meet the deadline requirements of messages. At the same time, the residual energy of the node and the average delay of the network are considered. In order to obtain as many messages as possible in parallel transmission, the message graph is established according to the parallel optimization model of message scheduling for the messages that need to be transmitted in the same time slot, and the multi-round conditional coloring of message graph is carried out by using graph coloring theory. Minimize the number of colors used, and messages corresponding to the vertices of the same color can be transmitted in parallel. Experimental results show that the ERMPSA algorithm based on the message scheduling parallel optimization model can maximize the degree of parallelism. Compared with the existing message scheduling methods, the ERMPSA algorithm can effectively reduce the loss rate of the message deadline without increasing the node energy consumption, and also has a better performance in reducing the average transmission delay of messages and improving the network throughput.
【学位授予单位】：南昌航空大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP29

【参考文献】