Traffic jam

By Lynley Hargreaves 24/04/2014


When Associate Professor Ilze Ziedins goes to work, she chooses between taking the ferry or driving her car. The car trip sometimes takes a mere 15 minutes, but if the roads are congested it can last an hour or more. So she has a personal interest in her research on the mathematics of networks, and how real-time information can help.

Associate Professor Ilze Ziedins in front of an uncongested section of State Highway One.
Associate Professor Ilze Ziedins near an uncongested section of State Highway One.

Are you out on the street measuring traffic queues and asking people about their travel choices?

No, the work we do is just with mathematical models of networks and idealized users. My background is in modelling queueing networks including, for instance, telecommunication networks, and the flow of patients through treatment centres such as intensive care units. For this roading project we ask, if people can choose the route that minimizes their delay, which route will they take, and how does this affect the congestion in the network as a whole? One of the questions that we find particularly interesting is how the balance between using cars and taking public transport (buses, trains, ferries) may change in response to information about current congestion.

How much does it help, in your models, for users to have real-time rather than average or historic traffic information?

This depends partly on the traffic load on the system. When the system is lightly loaded, information makes little difference. When the system is heavily loaded and all routes are congested, the primary effect may be through individuals cancelling or delaying their travel. But the biggest effect is at intermediate traffic levels where having real-time information can make a considerable difference simply through the choice of route, or travel mode (car vs. public transport). We all know that extraordinary events such as accidents, or very bad weather, will make a difference to congestion, but even without those there will be variation in flow from day to day, and therefore variation in congestion from day to day. With real-time information we can respond to that variation.

Is there actually enough traffic information out there to be useful in real life?

When I first started thinking about these problems real time information was only available on the radio in summary broadcasts. Now we have services such as the traffic feature on google maps, which gives both predicted congestion at any time of day and actual current congestion. NZTA also has web cams that monitor congestion at various places around the country. For instance, last week I was staying in Lower Hutt and was planning to go into central Wellington. A quick look on google maps showed that a combination of the royal visit, and bad weather from the tail of Cyclone Ita, had led to very heavy congestion all the way into town and in the CBD. I postponed the trip into town, had a delicious café breakfast in the Hutt, and wrote this blog instead! My decision led to only slightly less congestion (one less car), but if enough people make a similar decision then jointly that can have a considerable effect.

It sounds like we should all be looking at traffic cams, to make better travel choices. But can your work tell us anything about how to make the transport system itself better? Does building more roads keep congestion down?

Adding extra capacity to road networks is one way of reducing congestion, but if it isn’t done carefully, it may not have as much of an effect as expected. Partly this is because people may simply choose to take more trips. It’s also because of different route choices, or because people then choose to use public transport less. In the Downs-Thompson paradox, extra roading capacity means so many people may switch away from using public transport that the frequency of the public service may even be reduced, and everyone ends up worse off than before. Providing effective public transport for large cities is essential to combat congestion, and can help to reduce trips made by car. The striking reduction in congestion during school holidays illustrates how reducing car trips helps to reduce congestion – every time we take our cars out we are contributing to the congestion experienced by other people!

You can help reduce congestion by checking the traffic conditions before you travel at NZTA or by enabling the traffic feature on google maps.

These interviews showcase researchers supported by the Marsden Fund which, since 1994, has been supporting fundamental, investigator-led research in New Zealand.