计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (7): 30-34.

• 博士论坛 • 上一篇    下一篇

多功能车辆总线周期扫描表优化设计

曾秋芬,陈特放   

  1. 中南大学 信息科学与工程学院,长沙 410075
  • 出版日期:2015-04-01 发布日期:2015-03-31

Optimization design of MVB periodic scanning table based on simulated annealing algorithm

ZENG Qiufen, CHEN Tefang   

  1. School of Information Science and Engineering, Central South University, Changsha 410075, China
  • Online:2015-04-01 Published:2015-03-31

摘要: 针对参与通信的各周期信息特点设计优化的多功能车辆总线周期扫描表对提高列车通信网络的实时性能具有重要作用;标准推荐的周期轮询算法存在周期信息分布不均匀,带宽利用率相差较大的问题,在对周期扫描表的优化设计方法进行深入研究后,提出基于模拟退火算法的优化方案。建立了周期信息通信模型,详细阐述了通信抖动的概念并对其进行了定义。给出了优化目标函数并采用模拟退火算法进行求解,对算法的几个关键步骤进行了设计,并采用改进策略提高了模拟退火算法效率。通过实例仿真证明该算法实现了周期信息在整个宏周期范围内的均匀分布,优化了周期扫描表的构造。

关键词: 多功能车辆总线, 周期扫描表, 通信抖动, 单调速率算法, 模拟退火算法

Abstract: According to the characteristics of periodic messages, an optimized design of Multifunction Vehicle Bus(MVB) periodic scanning table plays an important role in improving the real-time performance of Train Communication Network(TCN). The periodic polling algorithm introduced by the standard has some drawbacks, such as uneven distribution of periodic messages and large difference between the utilizations of MVB bandwidth. After analyzing some optimized algorithms, a new optimized strategy is proposed based on simulated annealing algorithm. The periodic communication model is constructed. The concept of communication jitter is elaborated and a definition is made concerning communication jitter. The optimization objective function is given and can be solved by adopting simulated annealing algorithm. Some key procedures are designed. Some improved strategies are proposed for enhancing the efficiency of simulated annealing algorithm. Finally, the simulation results prove that the uniform distribution of all periodic messages is implemented for the range of the whole macro-period, and the establishment of MVB periodic scanning table is optimized.

Key words: Multifunction Vehicle Bus(MVB), periodic scanning table, communication jitter, Rate Monotonic(RM) algorithm, Simulated Annealing(SA) algorithm