智能交通中的动态径规划研究.docVIP

摘 要 最优路径规划是人们日常出行都要面对的问题，是智能交通的重要应用领域。本文以镇江市部分区域作为路径规划的研究对象，在MapInfo上建立了该区域的路网模型，并使用MapBasic进行二次开发，实现路网的拓扑，为路径算法的研究提供实验平台。 路网模型包括了路段的长度、连接关系等静态信息，基于这些静态信息的路径规划问题，传统的路径算法已经可以解决，其中Floyd算法求得的解是最完整的。本文通过加权路网模型的矩阵迭代实验，证明Floyd算法求出的最小权值矩阵包含了路网中所有顶点之间的最短路径长度。然后用Floyd算法对镇江市部分区域路网模型进行最优路径规划实验，将实验结果与导航软件进行对比，证明算法的正确性。最后将算法的运行结果作为蚁群算法研究的理论最优路径。 蚁群算法通过模仿蚂蚁的觅食行为来求出两点之间的最优路径，具有较强的灵活性，可以方便地与其他算法进行结合，但也存在容易过早收敛、陷入局部最优解等缺点。为此本文对蚁群参数进行实验调试，得出每个参数对算法性能的影响，给出合理的配置范围。针对算法存在的缺点，对算法的信息素浓度范围和更新方式以及概率选择方式进行改进，提高了算法的搜索性能。 根据Floyd算法和蚁群算法的特点，本文提出将两者相结合形成F蚁群算法，使得改进后的F蚁群智能化，可以模仿日常生活中人类的出行，实现交通拥堵模拟。在每个动态模拟周期使用改进蚁群算法进行最优路径规划。动态路径规划实验表明，改进后的蚁群算法可以根据实时路况及时调整出行策略，规划出避免交通拥堵的最优路径，对解决交通拥堵问题有一定的帮助。 关键词 最优路径；Floyd算法；蚁群算法；F蚁群算法；交通拥堵 Abstract Optimal path program is a problem people certainly meet with when going out and it is an important application area in ITS. This paper uses partial region of Zhenjiang City as research object for path program, set up road net model of this area on Mapinfo and uses MapBasic for secondary development to realize road net topology and establish experimental platform for the research of path algorithm. Road net model includes all the static information such as length of roads and topological relation and so on. Path program problems based on static information can be solved by traditional path algorithm, Floyd algorithm can work out the most complete solutions among these algorithm. Through the experiment of matrix iteration on weighted network model, this paper proves the minimum weight matrix calculated by Floyd algorithm includes the length between all the nodes in network. Afterwards path program experiment is conducted on the road net model of partial region of Zhenjiang City by Floyd algorithm and the experimental result is compared with navigation software to prove the validity of Floyd algorithm. The computational result is used as the theoretical value for ACO. ACO can get the optimal path between two nodes through imitating ant’s foraging behavior. It is very flex