考虑网络拥堵与系统公平的车载异构网络选择方法

利用演化博弈的有限理智特性执行网络选择, 以实现车载异构网络系统中网络资源的均衡分配; 利用双层博弈对演化博弈方法进行优化, 保证极端拥堵中部分车辆消息传输的同时, 维持系统公平; 设计了专用短程通信、长期演进和无线局域网融合的车载异构网络仿真场景, 对比了基于多准则决策的传统方法、基于演化博弈的网络选择方法和基于双层博弈的网络选择方法。仿真结果表明: 采用基于演化博弈和双层博弈的车载异构网络选择方法首次解决了动态网络环境中车载异构网络切换时出现的大规模乒乓效应, 利用双层博弈能够实现拥堵抑制和系统公平; 采用基于双层博弈的网络选择方法能够驱动异构网络系统在2~3个切换周期内实现网络系统状态的稳定; 在预设的动态网络评价条件下与80个终端的一般场景中, 双层博弈终端平均网络评价指标高于演化博弈19.5%, 为3种网络协同工作提供可靠服务; 在190个终端极端拥堵场景中, 终端合理分配, 共享专用短程通信网络资源, 双层博弈终端平均网络评价指标高于演化博弈10.3%, 双层博弈专用短程通信网络评价指标为演化博弈的2.18倍, 可以保证车联网基本安全信息的广播、系统的公平并维系基本车联网服务。 

Heterogeneous vehicular network selection method considering network congestion and system fairness

The bounded rationality characteristic of evolutionary game was used to implement the network selection, and the network resource of heterogeneous vehicular network system was evenly distributed. The fairness of the system was guaranteed by optimizing evolutionary game with two-layer game while some of the vehicles can transmit message in extreme congestion. The network simulation scene combined with dedicated short-range communication (DSRC), long-term evolution, and wireless local area network was designed, and the traditional method based on multiple criteria decision making, the network selection method based on evolutionary game, and the network selection method based on two-layer game were compared. Simulation result shows that the large scale ping-pong effects of dynamic network environment in heterogeneous vehicular network switching was firstly solved by using the heterogeneous vehicular network selection method based on evolutionary game and two-layer game. The two-layer game can suppress congestion and provide system fairness. The network selection method based on two-layer game can drive the heterogeneous network system to achieve the stability of network system state in 2-3 switching cycles. In the preset dynamic network evaluation condition and the general scene with 80 terminals, the terminal average network evaluation index of two-layer game is 19.5% higher than that of evolutionary game, and it provides reliable services for three kinds of network cooperation. In the extreme congestion scenario with 190 terminals, the terminals are reasonably distributed and share the DSRC network resources. The terminal average network evaluation index of two-layer game is 10.3% higher than that of evolutionary game, and the evaluation index of two-layer game DSRC network is 2.18 times of evolutionary game. Therefore, the basic safety messages broadcast, system fairness and basic connected vehicle service can be ensured.