Public Transportation Planning and Operations
Transit network Design
We developed a web-based "4-step model" of trip generation/attraction (hexagon colors), distribution, and assignment. The tool allows for adding and adjusting a transit network following the standard GTFS structure consisting of bus routes, blocks, and trips. The assignment step captures congestion levels represented here by the thickness of the lines. Transfers are allowed when the two related bus stops are in the same hexagon. The tool allows for adjusting features such as block headways and station locations. Next, we plan to implement the model on real networks and add perscriptive analytics and GTFS export capability.
Shuttle Service Design
Fundamental to Transit Planning is the definition of, and the relationships among, headway, load factor, fleet size, and tour duration. The demo shows the impact of (moving) one station on the crowdedness of the shuttle service, represented by the thickness of the lines. The performance metrics are also quantified in the bottom panel for further analysis.
Real-Time GTFS Analysis
Real-time GTFS feeds of bus locations come with a lag of 30 seconds (the public version), which makes their animation look like delayed stop motion. We developed a location prediction model for the buses between data refreshes, which makes the animation much smoother and comprehensible.
Bus bunching happens when two or more buses that are supposed to be evenly spaced end up arriving at a stop close together, like a "bunch." This happens when a delayed bus picks up more passengers and gets further behind, while the next bus has fewer stops and catches up. It leads to long wait times, overcrowded buses, and unreliability.
We wrote an algorithm based on GTFS-RT and its static version to detect bunching. The heatmap presents bunching intensity in the last 5, 10, 15, and 30 minutes. We will next implement optimal real-time bus-holding strategies to decouple bunched buses when it improves the level of service.
Transit Fleet Electrification
The many benefits of fleet electrification are driving transit agencies to electrify their fleets. The challenge is to choose an ideal fleet composition and size, deploy chargers at strategic locations with certain outlet capacities, and configure an optimal charging schedule that adheres to operational and regulatory constraints.
E-bus Pro is an interactive tool for transit planners to design electrification plans by choosing charger locations, outlet capacity, type, and electric bus specifications like the range and minimum required safety charge. The tool takes schedule data from the GTFS and generates an optimal en-route charging plan that adheres to scheduling constraints and other regulatory conditions.
In the background of E-bus Pro are optimization/prediction models that take from GTFS data the schedule of each bus—defined formally as a “bus block” in Transit terminology. The optimization maximizes a specific objective, subject to operational constraints, such as charge continuity and layover time restrictions.