Traffic is Just a Game!
This article briefly praises Samah El-Tantawy in Toronto for her application of game theory to traffic lights. Essentially traffic lights were given an AI that was trying to “win” at getting people through traffic. Each traffic light is an agent, which relays information to a main server. The main server then judges possible outcomes and chooses as best it can. (The algorithm is self teaching and learns how to make better decisions over time.)
This is a great example of how game theory can help in more than just games. By allowing traffic lights to monitor traffic the highly complex traffic systems in a big city like Toronto, traffic times could be increased almost 25%. Essentially the traffic system is aware of both scoring ‘goals’ or moving large swathes of traffic and ‘winning’ or getting everybody through the system. Then the system judges with a complete picture.
In a case such as police interrogator model, the algorithm knows to choose what is in the accompanying info graphic the Lie, Lie option, which traditionally is hard for people to choose on their own even though it makes it is the best option. This shows how constraining a game in this case with traffic laws and a server making decisions based on numbers. Such a system bypasses the dominant strategies of individual players and is ideally able to make the best decision for everyone. In this case, decreasing traffic.