You may be interested in this: http://sumo.sf.net (SUMO: Simulation of Urban MObility)
Ross
Interesting work Brad – I hope it leads to something useful. Tell, me, can you simulate all of the driver types we have in Auckland? I'm thinking in particular of the immigrants who either bought their licence before they came to New Zealand, or perhaps never even bothered to get one at all… But what really bugs me is how badly the traffic lights work here. I have moaned about this for years, but just how bad it really is was brought home to me when I went to Europe for the first time last Christmas. I didn't do a lot of urban driving, but one day spent driving around Geneva was truly eye-opening. The lights are designed and operated in order to assist traffic flow, not to disrupt it. The lights will typically be red on approach to an intersection, but time after time after time they turn orange then green just as you get there. Driving from one side of the city to the other, I had to stop only two or three times. The rest of the time I only had to slow to 40 km/h – if I had to slow at all. Auckland lights seem hell-bent on bringing every single vehicle to a complete halt – even if no other vehicle is using the intersection (or even if there is no other vehicle in sight), and even in the dead of night. I would like to think that someone could fix the problems we have here. You could easily make a serious dent in the time taken to drive around Auckland, saving big money for everyone. And if you add up the combined kinetic energy of our vehicle fleet and convert that to litres of fuel, the cost of even a couple of extra stops each day for every vehicle in the city works out to many thousands of dollars. To me the waste of time, money, and energy is nothing short of criminal and I have no idea why the people responsible for the system can't see it that way. Perhaps that SUMO simulation of Cologne uses a similar traffic light programme to Geneva? Ross McCorquodale
http://www.brad.net.nz Brad Heap
Interesting work Brad – I hope it leads to something useful.
At the moment I am trying to get a PhD scholarship to extend the ideas involved.
Tell, me, can you simulate all of the driver types we have in Auckland?
Each car in the network has its own randomly distributed normal variable that determines how close it sticks to the speed limit. A value of less than 0 will result in it driving slower than the speed limit, a value greater than 0 will see it exceed the speed limit.
But what really bugs me is how badly the traffic lights work here. I have moaned about this for years, but just how bad it really is was brought home to me when I went to Europe for the first time last Christmas. I didn't do a lot of urban driving, but one day spent driving around Geneva was truly eye-opening. The lights are designed and operated in order to assist traffic flow, not to disrupt it. The lights will typically be red on approach to an intersection, but time after time after time they turn orange then green just as you get there. Driving from one side of the city to the other, I had to stop only two or three times. The rest of the time I only had to slow to 40 km/h – if I had to slow at all. Auckland lights seem hell-bent on bringing every single vehicle to a complete halt – even if no other vehicle is using the intersection (or even if there is no other vehicle in sight), and even in the dead of night.
The simulator uses fixed green-wave traffic light controls. Which do not take into account the volumes of traffic flowing on the network. This is an option for expansion but people have done whole studies purely on traffic light controllers.
I would like to think that someone could fix the problems we have here. You could easily make a serious dent in the time taken to drive around Auckland, saving big money for everyone. And if you add up the combined kinetic energy of our vehicle fleet and convert that to litres of fuel, the cost of even a couple of extra stops each day for every vehicle in the city works out to many thousands of dollars. To me the waste of time, money, and energy is nothing short of criminal and I have no idea why the people responsible for the system can't see it that way.
Yes absolutely agree, but two issues: a) Full scale commercial simulations cost a lot of money – sometimes more than the amount of money they would save. b) Get the simulation right is very difficult as there are so many variables at play, hence why it is a complex system.
