【作者】 谭惠丽；

【导师】 孔令江； 刘慕仁；

【作者基本信息】 广西师范大学 ， 理论物理， 2002， 硕士

【摘要】 元胞自动机（cellular automata，简称CA）是空间、时间和状态都离散的非线性动力系统，在计算上具有并行性和简单性，易于在计算机上操作，用它模拟各种复杂现象已取得了很好的结果，近年来已被广泛应用于交通问题的研究中，是模拟交通现象的有效工具。 本文从交通问题研究的重要性和复杂性出发，在了解交通流的基本概念和现象的基础上，介绍了交通流理论的发展概况：从早期的车辆跟驰模型到当今的交通流元胞自动机模型，对各时期的主要模型及理论发展都作了介绍，并重点阐述了元胞自动机模型．本文以元胞自动机理论为基础，根据已有的交通流CA模型建立几个新的模型，研究它们的交通流特性。 首先，建立车辆只有0，2速度的单车道一维确定性交通流模型（称（0，2）模型），通过计算机模拟，所得结果与V_max=2的确定性FI模型（称（0，1，2）模型）进行比较，发现（0，2）模型的临界密度小于（0，1，2）模型的，而达到稳定态的时间则大于后者，由于车道上车辆间的相互作用不同，（0，2）模型较易于形成局部阻塞。 在单车道模型的基础上，建立V_max=2双车道模型，研究车道的耦合效应。双车道模型中车辆可以进行转道、超车，两道之间有耦合作用。模拟结果说明，有耦合时系统的流量比没有耦合时大；在固定的初密度条件下，只要耦合系数b＞0，两道间的耦合强弱对流量和速度无影响；两道初密度分布不均匀和速度分布不均匀都能使交通系统特性发生变化。 接着我们研究改进的Negal-Schreckenberg（NS）模型在开放边界条件下的交通流性质。改进的NS模型中，车辆在进行速度和位置更新时，除了和NS模型一样要考虑该车与前方车辆间的距离外，还考虑车辆运动过程中的相对位置。开放条件下，系统始端以概率α产生车辆，系统终端上的车辆以概率β离开系统。我们主要研究了系统车辆密度、平均速度和流量受边界开放程度、最大速度V_max和随机减速概率P的影响。考虑相对运动后，结果与开放边界条件下的NS模型结果相比，流量、速度均有提高，系统从自由运动相转化为最大流量相或从阻塞相转化为自由运动相的相变行为及其临界点都有所不同，另外还出现了一些新的特性。 最后，在开放边界条件下改进NS模型的基础上，建立二维城市主干道模型，研究不同参数条件下模型的性质，找出影响主干道流量、速度的因素及城市交通流变化的规律，给出适当的参数匹配，为主干道管理和控制提供依据。主干道模型由一条东西向的主干道和若干条南北向的支道组成，在支道和主干道的交叉路口上设置红绿灯，交叉口上的车辆可以转向行驶。通过计算机模拟，分别讨论了主干道模型的自组织和组织作用，其中主干道、支道的长度，交叉口数目及其位置的变化属于自组织作用，而减速概率P、产生概率Q、消大概率二、红绿灯周期可由人实时控制，属组织作用。我们分析了以上各种因素在城市主干道交通中所起的作用，了解了各种参数的性质后，就可以按实际情况选取合适的参数对各种情况进行模拟，预测交通状况。更多还原

【Abstract】 Cellular automata(CA)are mathematical idealizations of dynamical systems in which space,timea and states are dicrete.It’s parallel and simple for computation,so it can simulate variable complex phenomena.Resently,CA have become used for modeling traffic flow.In this paper,firstly we point out the importance and complexity of traffic problems.Then we make a summary of traffic flow theory:from the early car-following model to resently CA model,important models and their theory in diffrent period are presented,especially the CA models.Basing on the traffic flow CA models,we present several new models and study their traffic characters in the following.First,it’s a deterministic CA model for single lane.In this model,vehicles have the speed 0 and 2,so it is called(0,2)model.This model is compared with the deterministic FI model for Vmax=2 which is called(0,l,2)model.It founds out that the critical density of(0,2)model is less than that of(0,1,2)model,and the time to get to steady state for(0,2)model is longer,and it jams easily.Second,we introduce a two-lane traffic model which is based on the(0,2)and(0,1,2)models.We mainly study the coupling effect of lanes.In two-lanetraffic ,vehicles can change lanes and overpass.The interactin between lanes is considered as coupling effect.The computer simulation indicate that(Dthecurrent of the system with cupling is greater than that without coupling;0for a certain initial density,the average flow and speed are indepent on the coupling coefficient which is greater than zero;when the initial densities(speeds)of lanes are different,the characters of traffic system will be different.Third,we investigate the improved NS model with open boundary conditions.Based on the NS model,this model considers the relative motion of vehicles when updating the speeds and positions of vehicles.With open boundary conditions,at site i=0(out of system),a vehicle with probability a and with velocity V=Vmax,and at the end of system,vehicles simply move out of the system with probability 3 .The influence of a 3 Vmax and deceleration probability P on the car density average speed and current are investigated.Compared with NS model,the quantities in our model such as density,speed and current are improved greatly.The phase transition from free-flow phase to maximum phase or from jamming phase to free-flow phase is different from that of NS model and some new characters occure.Finally,we propose a main road model in city traffic according to the improved NS model with open boundary conditions,and study some characters about this model with different parameters.Our purpose is to find out the actor which influent the main road’s average speed and current,and also the law of city traffic.Then we can give some advise for managing and controlling the main road trffic.Organization and self-organization cooperation in this model is discussed.The effect of roads’length,the number of intercections and their positions is self-organization,and the effect of the deceleration probability,the injection(extinction)rate,the cycle times of traffic lights is organization.We analyse the effect of the parameters above separately.更多还原