A spatiotemporal data model for dynamic transit networks

A transit network is a spatiotemporally dynamic system whose topology, components, and services all change over time. The complexity and dynamic nature of transit networks make it difficult to model by traditional transportation models. This paper presents an object-oriented data model that coherently represents space, time, and dynamics of transit networks and supports schedule-based pathfinding. In this data model, a transit network and its components are modelled as spatiotemporal objects that have properties and behaviours. Unlike traditional roadway transportation network models that define topology as connectivity between geometric elements such as edges and nodes, this data model uniquely defines topology as sequences and collections of transit features which are application domain objects. Moreover, this data model supports the generation of active network topologies based on given times to represent real-time network states as well as to reduce topological redundancy. Implementations of the data model in two Internet transit trip planners have demonstrated that it can effectively reduce computation complexity of schedule-based pathfinding algorithms and enhance pathfinding performance.