DEMAND-DRIVEN FORMATION OF TRANSPORT NETWORKS: EMERGENCE OF HUB AND SPOKE STRUCTURE

Abstract

Many transportation networks, including rail, sea, and air, have a characteristic structure known as hub-and-spoke (HS). Most previous studies have explained the formation of HS structure through the supply-side behavior without carefully considering effects of demand-side behavior on network formations. This study reconsiders Mori (2012) hub formation model from a gametheoretic perspective in discrete space and extends it to a transport network formation game. This game can describe the emergence of the HS structure from the demand-side behavior. In this game, agents choose travel links (transport services) while considering the link cost, which benefits from economies of scale. By agents’ behavior, a transportation network is formed as the equilibrium state of the game. We analyze the properties of the equilibrium state and clarify the emergence mechanism and stability of the HS structure in transport networks. The keys of our analysis are that the proposed game is a potential game and that the corresponding potential maximization problem has the structure of a Minimum Concave-cost Network Flow Problem (MCNFP). Based on the facts, we introduce a global optimization algorithm designed for MCNFP and derive the stochastic stable state as the global optimal solution to the potential maximization problem. Our numerical result demonstrates that the transport network with the HS structure can be formed as a stochastic stable state and that the location and spacing of hubs vary depending on how economies of scale work.