Improving probabilistic traffic modelling through advanced sampling

In traffic models certain presumptions are often made to simplify the complex systems that rule the world of traffic flow. This is necessary as not every variable can be considered. Furthermore, it is commonplace that equilibrium states are sought that give a good average, or rather deterministic, representation of the dynamics of traffic. Such an approach makes presumptions of traffic demand and supply for a (non-existent) average situation. However there must be a realisation that traffic is hardly ever ‘average’ [1]. It is especially in the terms ‘average’ and ‘deterministic’ that a realisation must exist that these terms are composed of extensively varying situations. By considering real stochasticity in these processes, a more complete picture of the traffic system is gaine