Using Games to Further Research



Transportation students gathered in the ITS Library Tuesday to have a little fun and play a traffic game that routed flows through a transportation network in the name of research.

The traffic game simulated a routing network, where players used the same network to send traffic to model drivers commuting on a road network, or Internet Service Providers sending traffic on the internet.

Players were given a source and destination on the network and tasked to send traffic between the points to minimize the latency of traffic with only a certain number of routes available. The game was played in rounds, where players choose how much traffic to allocate to each route. At the end of the round, players observed the latency of each route, and could change the allocation on future rounds.

Complicating the game, other player’s decisions affected the latency of routes, so routes that had more traffic were more congested, thus had more latency. However, players did not directly observe the decisions of other players, only the latency of their routes.

The goal of the experiment is two-fold:

(i) Researchers evaluated the capacity of humans to reach an equilibrium in a distributed environment (that is, without observing the decisions of other players). An equilibrium is a situation in which each player is no longer able to reduce his/her latency, assuming all other players do not change their decisions.

(ii) Researchers studied the decision-making process of the players, and develop a mathematical model of it, in order to understand how humans adapt their decisions to a changing environment.