Lint Test Coverage: total Maintainability
A TypeScript implementation of RAPTOR algorithm. It includes the following derivatives:
- RAPTOR
- McRAPTOR
All optimizations are included except local pruning ; see this issue. It is implemented on the local-pruning branch
An algorithm to compute all Pareto-optimal journeys in a dynamic public transit network for multiple criteria, such as arrival time and number of transfers.
Implemented criteria at src/criteria.ts:
- Foot distance, minimizes the foot distance over a journey
- Buffer time (same as the original paper), maximizes the minimum transfer time to hop on a trip over a journey
- Success probability, relative to how connection times intersect
Some new features are brought in this implementation:
- Enhancing Pareto front to be maximal in general (with any criterion) — issue, PR
This is done by scanning subsequent catchable trips, not only the earliest one. The scanning stops when its label is dominated by one of a previously scanned trip. This implies the Pareto front is maximal if every criterion is increasing for increasing trips. - Generalizing the time type. It can be a scalar, an interval.
- Ability to run multiple queries (multiple RAPTOR instances, in fact) in parallel, reading the same RAPTOR data (memory chunk). This is done thanks to a SharedArrayBuffer together with powerful views. There is a little overhead due to the views.
- Daniel Delling, Thomas Pajor, Renato F. Werneck (2014) Round-Based Public Transit Routing. Transportation Science 49(3):591-604. doi.org/10.1287/trsc.2014.0534