Gravity model for spatial network based on optimal expected traffic

Abstract

In this paper, we propose a spatial network model based on the optimal expected traffic. The model maximizes the total expected traffic of the network. By changing two parameters α and γ which represent the dependence of the network on the fitness and geographical constraints respectively, the model can vary its topology from heterogeneous star-like networks to homogeneous road-like networks. We use our model to simulate the Chinese city airline network and compare it with the real data. In the end of the paper, we discuss the relationship of the expected and the real traffic （namely the real weight on the network）.