F. Combes, A. De Palma
Consider a population of agents who choose one among a set of destinations located along a rectilinear road. Each of these destinations has a certain utility, modeled by a random variable. We compare a situation where the agents have to explore the destinations to observe the value of their utilities, to a situation where they know these values beforehand. We show that more information yields a higher welfare to the agents, and also, perhaps counter-intuitively, higher distance traveled.