Measuring Stability of Stochastic Model of Traffic Flow
Ahmad Mahmoodjanlou, Martin Hazelton, and Katharina Parry
Institute of Fundamental Sciences, Massey University, Palmerston North
In this talk, we look at scenarios where some unforeseen occurrence happens to a traffic network on a particular day. For example, temporary road closures or an accident disrupts the normal flow of traffic. We would observe that on such days the cost of using affected routes increases, given that the incident would mean it takes a long time to pass through that route. Using a measure called the coefficient of reactivity, we investigate the how fast a system returns to normal after various types of such one-off events. We then show how this measure can be used to determine in which systems a stochastic model of traffic flow in can be well approximated by its deterministic counterpart.
This presentation is eligible for the NZSA Student Prize.