An Evolutionary Model for Efficient Transportation Networks

In this paper, we present a model to automatically generate efficient transportation networks given a simulated urban environment with predefined population distributions and other physical constraints. Based on the empirical analysis of different topological structures of networks, we found that that the efficiency of transportation networks heavily depends on the layout of the stations. The model uses the genetic algorithm to optimize the spatial distribution of stations. Then, the Minimum Spanning Tree is constructed upon which extra network routes are built to minimize average travel time within the constraint of a network length. Experimental studies of a simulated Beijing subway system showed that our model can generate a network that is 14% more efficient than the current subway system based on the Beijing population distribution and geography. This study can be extended to other transportation systems designs, and other communication networks.