【作者】 张欣；

【导师】 刘辉林；

【作者基本信息】 东北大学 ， 计算机系统结构， 2015， 硕士

【摘要】 车载自组网基站建设作为国际上公认的解决城市交通拥堵问题最经济有效的办法之一,受到社会各界的广泛关注。基站节点覆盖问题是车载自组网的关键技术之一,在城市道路交通流的均衡动态分配中起着举足轻重的作用。本文正是基于这样的应用背景,针对基站节点部署的核心组成部分:车载自组网络中服务基站扩展覆盖算法进行了基于混合微粒群算法的研究。文章内容主要包括以下两个部分:基站扩展覆盖模型的建立以及混合微粒群算法的设计。第一部分:对城市交通路网进行抽象描述以后,针对城市交通路网的特点,以及车辆节点移动的动态性,通过理论分析得出服务基站节点扩展覆盖某一区域的概率与服务基站扩展覆盖半径的关系,结合实际情况通过设定扩展覆盖概率值大于某一特定值,得出覆盖半径的最小值。第二部分:根据微粒群算法的不足,本文提出了一种基于扩散机制的双种群微粒群优化算法。主要在三方面进行了改进:为提高算法后期的全局寻优能力和保证微粒群物种的多样性引入了多种群思想;为提高算法后期的收敛速度和寻优精度,引入了扩散机制;为保证算法迭代计算的高效进行,增加扩散池来去除对全局寻优能力贡献弱小的微粒并实现多种群信息的交流和扩散,避免陷于局部最优。最后,木文通过仿真实验验证了改进后模型和算法的有效性,且在求解效率和求解精度方面均要优于基本的微粒群算法和传统的遗传算法。更多还原

【Abstract】 Vehicle network base station construction,which is internationally recognized as one of the most economic and effective way to solve urban traffic congestion problems,received widespread attention in the community.The node coverage problem is one of the key technologies in Vehicular Ad Hoc Networks,plays an important role in balancing the dynamic allocation of urban road traffic flow.This thesis is mainly based on backgrounds as mentioned and aims at Service station extended coverage algorithm research based on Hybrid Particle Swarm Algorithm,which is the hard core of the base station node deployment components.The first part:After the abstract description of city road network,according to the characteristic of city traffic,and the dynamic vehicle mobile nodes,The theoretical analysis shows that the relation between service base station node coverage probability and extended service station of a certain region of extended coverage radius,combined with the actual situation by setting extended coverage probability value greater than a specific value,it is concluded that the covering radius of minimum.The second part:According to the deficiency of the particle swarm algorithm,this paper proposes a double population particle swarm optimization algorithm based on diffusion mechanism.The main improvements are in three aspects:in order to improve the algorithm at the end of the global searching ability and ensure the diversity of particle swarm species introduced a variety of group thinking;in order to improve the convergence rate of the algorithm for the late and the precision of the optimization,introduced the diffusion mechanism;in order to ensure the efficient iterative algorithm for computing the increased diffusion cell to remove the global search optimization ability with small particle and to achieve a variety of group information exchange and diffusion,avoid getting stuck in local optima.The final experimental results show that the validity of the improved model and algorithm is superior to the basic Artificial Fish Swarm Algorithm and traditional Genetic Algorithm from angle of efficiency and precision.更多还原