Contrary to Popular Belief: New Roads Can Increase Traffic Congestion
The explanation is a surprising paradox of Game Theory: New York City temporarily closed 42nd street in 1990 and traffic improved. After removing a motorway in Seoul, South Korea, the traffic around the city sped up. In 1969, a newly built road in Stuttgart, Germany, slowed traffic down and had to be closed again.
Santiago
Machine Learning. I run https://t.co/iZifcK7n47 and write @0xbnomial.
-
Contrary to popular belief:
— Santiago (@svpino) March 23, 2023
• New roads can increase traffic congestion.
• Closing roads can improve traffic.
The explanation is a surprising paradox of Game Theory: pic.twitter.com/kieWu6IjJQ -
New York City temporarily closed 42nd street in 1990. Traffic improved.
— Santiago (@svpino) March 23, 2023
After removing a motorway in Seoul, South Korea, the traffic around the city sped up.
In 1969, a newly built road in Stuttgart, Germany, slowed traffic down and had to be closed again.
What's going on? -
Breess's Paradox was discovered in 1968 by the German mathematician Dietrich Braess.
— Santiago (@svpino) March 23, 2023
Imagine a city with 2 roads connecting A and B.
It takes 60 minutes to go from A to B, regardless of the route you choose and how heavy the traffic is.
But there's a bridge on each road. pic.twitter.com/vdbE12jRwg -
The bridges are bottlenecks.
— Santiago (@svpino) March 23, 2023
The time it takes to cross these bridges varies in proportion to the rate at which cars travel them.
For instance, crossing a bridge takes a full hour if everyone uses the same road. 30 minutes if only half the city tries. -
Unsurprisingly, every morning, people will choose the route with less traffic or a random road if they don’t know better.
— Santiago (@svpino) March 23, 2023
This selfish approach leads to an optimal solution:
Half of the people will go left, half will go right, and everyone will spend around 90 minutes driving. -
But the authorities decide to build a new supersonic highway connecting both bridges.
— Santiago (@svpino) March 23, 2023
Imagine this is a magic highway, and it takes no time to cross it.
It should improve traffic, right?
Counterintuitively, it doesn't. pic.twitter.com/iNqbv6vZRL -
On opening day, everyone wants to take the highway to reduce their travel time to 60 minutes.
— Santiago (@svpino) March 23, 2023
Except, everyone makes the same calculation!
All the traffic migrates to the highway, causing the crossing time across the bridges to double. -
Before the highway, people spent 90 minutes in traffic. After, it took them 2 hours!
— Santiago (@svpino) March 23, 2023
One morning you decide to try your luck with a different route.
All three routes take the same time: 120 minutes, so you choose the left route.
What happens now? -
You still drove for 2 hours.
— Santiago (@svpino) March 23, 2023
But the traffic across the second bridge improved slightly because you didn’t take the highway.
Everyone except you benefited from your decision that day. -
But you’ll want to pick the least congested route the next morning.
— Santiago (@svpino) March 23, 2023
Everyone is now back on the highway with no incentive to switch! -
Before the new road, we had a Nash Equilibrium, and everyone spent 90 minutes on the road.
— Santiago (@svpino) March 23, 2023
After the new road, we have a new Nash Equilibrium, but it's now a worse one: 120 minutes in traffic.
By building an extra road, the authorities made traffic worse, not better. -
I originally wrote this story for my newsletter subscribers.
— Santiago (@svpino) March 23, 2023
Tomorrow, I'm sending them an introduction to the idea that turned searching for matching keywords antiquated overnight.
Subscribe here to receive it in your inbox:https://t.co/WrAVnRGoNM