摘要
首先建立交通流动力学模型求解问题Ⅰ.在不考虑流量和考虑流量的两种情况下,该模型都能够解出在任意给定的时刻t位于第一个传感器的车辆到达第5个感应器的行车时间.我们还从四个方面给出了判断交通堵塞的衡量标准,并且利用神经网络方法准确地对未来的车流状态进行了预测.问题Ⅱ建立了交通网络的加权有向图模型,引入协方差矩阵描述网络中道路之间的相关性,并设计了查找最优路径的动态Dijkstra算法.问题Ⅲ构建了统计多目标规划模型,利用车比雪夫不等式,成功找到了从端点3到14和14到3的最优路径,并估算出了对应的行车时间.
First, we set up a traffic kinetic model to solve the question Ⅰ . Not only taking no account of the flow but also considering the flow we all can get the answer of question Ⅰ . We can get the standard of estimating the traffic jam or not. And we also use the nerve net to forecast the intending freight flow well and truly.
Then we set up the directional traffic map model to solve the question Ⅱ . We introduce the coefficient matrix to describe the relativity of the roads in the net and design the dynamic Dijkstra arithmetic.
Finally we set up statistical several aiming design model to solve the question Ⅲ . Using the Chebyshev' s Inequality, we can find the best way from 3 to 14 and 14 to 3 and can also get the travel time.
出处
《数学的实践与认识》
CSCD
北大核心
2006年第7期22-30,共9页
Mathematics in Practice and Theory
关键词
交通流动力学
统计多目标规划
车比雪夫不等式
traffic kinetic model
statistical several aiming design model
the Chebyshev's Inequality