基于实时调度与图分割的汽车网络结构设计研究  

Research on Automobile Network Structure Design Based on Real-time Scheduling and Graph Partitioning

作　　者：王跃飞[1] 韩江洪[1] 张建军[1] 彭浩[1] 

机构地区：[1]合肥工业大学,合肥230009

出　　处：《中国机械工程》2010年第13期1623-1626,1637,共5页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(60873003);高等学校博士学科点专项科研基金资助项目(20070359028);广东省;教育部产学研结合项目(2008B090500226);合肥工业大学博士专项科研基金资助项目(GDBJ2010-008)

摘　　要：为满足多网段结构汽车网络设计需要,分析了网络中消息传输的调度机制,推导出了网络可调度判定条件;以该判定条件为网络实时性约束条件,把网络结构设计转化为子图划分问题,给出了汽车网络结构设计的优化模型;将小生境和自适应策略引入到遗传算法中,建立了该模型求解的小生境自适应遗传算法。实例应用证明了该方法的可行性,它不仅可以降低网段间数据传输的负载量,而且可以确保网络消息传输的实时性。To meet the requirements of designing automobile network with multiple segments,a judgment condition for schedulability was deduced based on analyzing message scheduling mechanism.An optimization model under this condition constraint was presented when network structure design was described as graph partitioning.To solve the optimization problem,a niched adaptive genetic algorithm was built by introducing niche and adaptive strategy.The applications indicate that the design method can reduce loads across different network segments,and guarantee network schedulability of messages in automobile network.

关 键 词：汽车网络 网络调度 图分割 自适应遗传算法 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]