Effect of congestion costs on shortest paths through complex networks.

We analyze analytically the effect of congestion costs within a physically relevant, yet exactly solvable, network model featuring central hubs. These costs lead to a competition between centralized and decentralized transport pathways. In stark contrast to conventional no-cost networks, there now exists an optimal number of connections to the central hub in order to minimize the shortest path. Our results shed light on an open problem in biology, informatics, and sociology, concerning the extent to which decentralized versus centralized design benefits real-world complex networks.