A test bench for evaluating spatial indexation methods for connecting points to large networks

This study deals with performance issues for connecting points to a network in geographic accessibility studies. The problem is analysed by the development of a test bench where the spatial indexation methods kD-tree, Morton code and fixed grid are implemented. Using the test bench scenarios of networks, point datasets, spatial indexation methods, and methods for computing distances between point and line segments are evaluated. A case study indicates that the indexation methods kD-tree and fixed grid have the best performance for connecting point to the network. Furthermore, the results show that an “algebraic method” is much faster than a “trigonometric method” for computing the shortest distance between the point and the network.