GPS making traffic worse?

In this article the author brings up the idea that everyone having a GPS, and thus up to date traffic information is actually making traffic worse. Apparently the time it takes you to find an ideal route with minimal traffic actually ends up causing traffic further down the network: "the researchers found that individuals waste up to 30%, 24%, and 28% of their travel time, respectively, due to choosing a personally optimal route instead of a socially optimal one." These percentages can be though of as the price of anarchy in this situation, i.e. how much society is missing because individuals act selfishly. This also related to Braess Paradox; how adding resources (GPS) can actually make traffic worse.

This is some really good thinking. Having GPS would assure that everyone would take their optimal route, and therefore reach a Nash equilibrium every time. However! If road and city planners knew the way people were going to go every time, they could properly plan how they would build roads to reduce traffic. That is, currently people have to rely on less concrete information, so they way they select roads or "play the game" is difficult to predict. But, if everyone has a GPS in the future, road planners will know exactly which option they would select. So, planners could build different roads depending on different destinations to increase social welfare. In a way, they could create a dominant strategy for each player by knowing their preferences. Although the paradox has definitely been proven, it is safe to say that if we could predict each person's actions absolutely, a central planner could maximize social welfare. The issue in the past has always been properly predicting these actions.

Alternatively, if GPS becomes regulated and widespread, it could actually be used to make people take the socially optimum routes. Think about it. Since a GPS monitor knows how many people are on each route, it could direct people along the route that yields a social optimum. It would direct x people one way, and y people along another. In a sense, it could act as an enforcer to direct people to the social optimum (not just watching people for speeding, but give different routes in an ideal manner). Because a third party director can regulate games to better social welfare than the Nash Equilibrium can in many situations, GPS could actually lower average driving times for everyone if some optimization was used.

OP forgot the article link: http://www.physorg.com/news142157939.html

I guess if you are solely using a GPS to optimize your driving experience at the moment, it's not a good idea since it is a computer calculating a different route based on current traffic hazards and road changes. Although it may seem that taking x road may be faster, these devices do not consider the traffic that goes on most roads.

GPSes are supposed to be devices that help you get from point A to point B. Getting there faster is just icing on the cake, so I would think it's not the GPSes are causing slowdown in traffic.

Philip Goins wrote:Alternatively, if GPS becomes regulated and widespread, it could actually be used to make people take the socially optimum routes. Think about it. Since a GPS monitor knows how many people are on each route, it could direct people along the route that yields a social optimum. It would direct x people one way, and y people along another. In a sense, it could act as an enforcer to direct people to the social optimum (not just watching people for speeding, but give different routes in an ideal manner). Because a third party director can regulate games to better social welfare than the Nash Equilibrium can in many situations, GPS could actually lower average driving times for everyone if some optimization was used.

This is a great idea, although it would need a lot of synchronization and an advanced data traffic network. Speaking of which, it would probably cost people money to use a feature like that already (i.e. data plan), and I would imagine the prices for these data plans are quite high, probably roughly $20/month. In that case, it probably can't be used enough to allow new entrants into this market to use this feature, and a cascade would fail.

The first routing example we discussed in class demonstrated that the routes that generated the social optimum were not the routes that individuals would actually take. In that example, in order to reach the optimum, half of the drivers in the network had to take one for the team, and take the back road even though the highway was faster.

If we had GPS devices that led us in certain directions in order to create a social optimization, then many of us would end up going a way that is good for society but not us as individuals. The result would be that nobody would trust the GPS, and we would all take the highway: which is what would have happened anyway. That was the whole purpose of the example: Greed creates a PoA that costs society.

But if the GPS was constantly updating itself with information about where there was traffic, the number of drivers taking certain routes would keep fluctatuing and the traffic network would stay in equilibrium. I might be interpreting this wrong, but go back to pages 130-134 in the text )chapter 7 where they discuss best-response dynamics and potential energy. Though I am not completely clear on what the authors are saying, I think that might change the perspective of some people.

Lauren Victory wrote:But if the GPS was constantly updating itself with information about where there was traffic, the number of drivers taking certain routes would keep fluctatuing and the traffic network would stay in equilibrium. I might be interpreting this wrong, but go back to pages 130-134 in the text )chapter 7 where they discuss best-response dynamics and potential energy. Though I am not completely clear on what the authors are saying, I think that might change the perspective of some people.

even if GPS were constantly updating itself, there is always a lag between your information and when you are actually at the road.

suppose we have 2 paths, one where time needed is 0.5 (road A) and the other is x (road B).

suppose if at t=0, 30% of the traffic is already on road B, and 70% is just about to make their decision.
GPS would advise taking road B because it is current faster than road A ( 0.5 vs. 0.3)

so at t=0+ or t=1, the 70% making decision will all goto road B, making everyone worse off.

I think using GPS in people's car to figure out the speed of traffic on a certain road is a great idea and one I hope GPSs start using in the future. I think having the ability to know how long it will take to travel on the road based on traffic at that moment is an idea that can definitely make people take more efficient routes.

One issue with it is the delay from the number of people on a road 10 miles ahead compared to when you actually arrive at that point can cause a problem. During that time a bunch of GPSs could tell all the cars to go one route because it is slightly faster and when everyone follows, it will cause a traffic jam. However, you could also make the GPSs coordinate with each other and figure out which percent of people it should tell to go one direction and which percent will go another way. So if one route has 40% of the people and another percent has 60%, the GPSs could tell 60% of the future people to go the first way and 40% go the second way and thus evening out the two routes.

We are also assuming the slow travel on the road is just caused by traffic. Some of it is also cause by road construction. This means the GPS could be used to allow drives to avoid slower roads that were caused by some external effect other than traffic.

One possible solution to this problem would be applying " Contrarian" principles seen in investing when interpreting GPS results and selecting routes.

Contrarian in investment basically means, when a good news or event projects the stock to go up, instead of buying, you do the opposite and sell. Vise versa for when a bad news hits the stock.

So when GPS tells you that route A is the most optimal, then it might be more efficient to go route B suing the Contrarian principle. or else you might find route A congested with cars by the time you arrive.

In South Korea, they started to make progress on making more efficient guided system. (This was able because of its comparably small size of Korea and market) All the GPS system measures the time per distance in certain area simulatenously and calculate the average time per road, and send information to the GPS for calculating sub-way for drivers. Still, there are many problems in this system and limits. Although, it uses average speed in road to decide the traffic, one can't truly tell the data is robust. Second, there is always delayed time.

However, it is making progress little bit and some day, it will achieve social optimum...