A Study Of Three-Lane Cellular Automata Traffic Flow Models

For both developing and developed countries, effective transportation systems are of primaryimportance in the development of society and economy. The establishment of efficient traffic models toreveal fundamental traffic characteristics is an essential requirement in the design, planning and operation oftransportation systems. In this field, cellular automata(CA) models have drawn increasingly more attentionin the study of traffic flow. Although the CA models are based on very simple sets of rules, and beingrealized with computer, they are remarkable in their ability to simulate both gross and subtle features of realtraffic. CA is nonlinear dynamical system in which space , time and state values are discrete . Using simplerules, CA has attained very good results in simulating several complex natural phenomena and physicalproblem. That enlightens people to simulate traffic flow by CA method. The dissertation focuses on the study of three-lane highway and the main contents are listed below. At first, based on theoretical models, the original NS model is applied and extended to three-lane CAtraffic flow models on a highway. These models satisfy periodic boundary conditions so that first and lastvehicles are identified, and the vehicles are effectively arranged on a circle. Secondly, a three-lane uncontrolled CA traffic flow model is proposed to describe the highway traffic:the model is symmetric with respect to the three lanes as well as with respect to the same type of vehicles.Numerical simulation is conducted and leads to some new results. The fundamental diagram obtained by thesimulation shows the traffic capacity of road is enhanced compared with that of the NS model with a singleVmax . The impacts of deceleration probability and lane-changing rate on the traffic flow and lane usage areinvestigated. Thirdly, considering the first lane is used for safe overtaking, a three-lane controlled CA traffic model isproposed: different rules govern the changing from the first lane to the second lane and the reverse process,so that the cars on the first lane always try to return to the second lane, independent of their situation on the广西师范大学硕士研究生学位论文first lane. The characteristic of traffic of the three-lane controlled traffic model is alike to that of thethree-lane uncontrolled traffic model．It turns out the three-lane controlled traffic model is helpful to enhancetaffic safety. Rourthly, based on studying the effect of introducing a certain fraction of vehicles with a lowermaximum velocity to the three-lane uncontrolled traffic model with a single Vmax , we propose a three-laneuncontrolled mixed CA traffic model where the vehicles are of two types characterized by two differentvalues of Vmax , namely,Vmax and Vmax corresponding to the fast and the slow vehicles, respectively. The f sresults are: for the same deceleration probability, at low densities, the contribution to the flow and the meanvelocity made by the fast and slow vehicles is higher than those of the NS single-lane model as well as thethree-lane uncontrolled traffic model consisting of only one type of vehicles with Vmax = 3, whereas lowerthan those with Vmax = 5; However, with the increasing in the total density, the flow and the mean velocityof the mixed vehicles are higher than those of the NS single-lane model; But in the high density regime,there is hardly any difference between the three conditions. At a cetain density region, various differentlane-changing rates have effect on traffic. With the increasing total density, the distribution of lane-changingfrequency presents two distinct maxima, which invites the interpretation of superposing two differentstructures. The corresponding maximum flow and maximum velocity in the model are lower than in thethree-lane uncontrolled traffic model by the fast vehicles, and the critical density is larger than that of thethree-lane uncontrolled traffic...