Public transit network optimization via a novel connectivity parameter

Abstract

The public transportation options in suburban North America are typically focused on commuting based routes toward a cities central business district. Routes often pass though suburban towns without connecting to points of interest or population centers.Additionally, North American suburban transportation routes are lacking first mile options with minimal local bus or rail services. This thesis tackles the local transit issue by developing a mathematical framework for the utility of each point in the region, this utilityis then aggregated into a connectivity metric. The connectivity metric improves upon standard public transit accessibility metrics by in corporating internal network distances.The connectivity metric was applied to real world problems by utilizing it as the objective value for the generation and optimization of a new local bus route. The mathematicalframework for the utility density can be adapted for any set of spatial data to be used as an input. For this thesis, the population density and points of interests were combined using travel survey data to define the utility density. The utility density value represents general transit attractiveness for all trip types. The connectivity metric was then computed for the region by computing the shortest path between each pair of points and scaling the travel time with the destination utility.With a well-defined connectivity metric, the town of Winchester, Massachusetts was used as a case study for a route optimization task. The utility density was computed using the mathematical framework and data from public data sources. The route optimization was completed using the generalized variable neighborhood search meta-heuristic.The optimization process returned a route that forms a distinct flower pedal shape centered at the Winchester Center regional rail station and the pedals reach out to different areas of the town. The optimal route was computed thirty times with an average connectivity improvement of 16.0%. The proposed route layout, effectively connects the town based on the optimization parameters and offers useful first mile options due to connections to all existing though town public transit lines.