Braess Paradox in the Laboratory: Experimental Study of Route Choice in Traffic Networks with Asymmetric Costs

The Braess paradox (BP) in traffic and communication networks is a powerful illustration of the possible counterintuitive implications of the Nash equilibrium solution. It shows that, paradoxically, when one or more links are added to a directed network with affine link cost functions that depend on congestion, and each user selfishly seeks her best possible route, then the equilibrium travel cost of each and every user may increase. We report the results of a traffic network game experiment designed to test the implications of the BP. The experiment included two network games: a basic network game with three alternative routes, and an augmented network game with two additional routes. Both networks included asymmetric link cost functions, and each game was iterated 60 times with complete outcome information. On each round of the game, the subjects were asked to independently choose a route from a common origin to a common destination in an attempt to minimize individual travel cost. Focusing on aggregate and individual frequencies of route choice and route switching, our results show that with experience in traversing the network, aggregate, but not individual, choice frequencies approach the user equilibrium solution as implied by the BP.