Abstract
Digital simulation studies have been carried out on many traffic phenomena because, being stochastic, they are frequently too complex to be represented by a reasonable mathematical model. Furthermore, empirical laws are difficult to obtain because con ditions on the road rarely remain stable for a sufficient time to obtain reliable data. Most simulations have been carried out to determine maximum flows, queue length, delays, etc. in terms of conflicting traffic flow, driver behaviour para meters, etc. A number of examples of this type of simulation are mentioned in the paper. The author's simulation program, however, was written to give information on vehicle behaviour within the platoon dispersion process. Having modelled starting condi tions at the stopline and arrival patterns at points along the road sufficiently well, it was assumed that vehicle behaviour within the process was also being accurately modelled. The queues of vehicles at stopline are of random length with randomly selected vehicle type and random performance para meters. Leading vehicles in each lane accelerate according to an exponential law, and other vehicles are given accelerations according to one or other of the car- following laws. A novel technique allows driver reaction time to be taken into account. Vehicles may turn from the main road and overtake, the simulation program giving data on every vehicle's position, speed, and acceleration at each scan. The paper describes in detail the more important of the steps which make up the whole complex dispersion process and the various measures taken to verify the simulation.

This publication has 3 references indexed in Scilit: