Faster Preprocessing for the Trip-Based Public Transit Routing Algorithm

We propose an additional preprocessing step for the Trip-Based Public Transit Routing algorithm, an exact state-of-the art algorithm for bi-criteria min cost path problems in public transit networks. This additional step reduces significantly the preprocessing time, while preserving the correctness and the computation times of the queries. We test our approach on three large scale networks and show that the improved preprocessing is compatible with frequent real-time updates, even on the larger data set. The experiments also indicate that it is possible, if preprocessing time is an issue, to use the proposed preprocessing step on its own to obtain already a significant reduction of the query times compared to the no pruning scenario