Parking Model#
The parking model in POLARIS simulates the end-to-end process of parking for personal vehicles, shared mobility fleet vehicles, and B2C/B2B freight vehicles within the urban transportation network. Parking is a critical component of the travel experience for auto-based trips: it affects route choice, destination choice, travel time, and cost. In POLARIS, parking is modeled as a set of interacting agent-based components that span the demand model (pre-trip parking choice), the traffic simulation (en-route parking search and vehicle parking/unparking), and fleet-specific operations.
The parking module captures the following key behaviors and processes:
Pre-trip parking choice: Before departing on a driving trip, person agents evaluate available parking options near their destination and make a choice between on-street and garage parking using a discrete choice model.
En-route parking search: While driving on the network, vehicles assess the urgency to park and search for available parking in real time, potentially rerouting to reach a parking location.
Parking facility operations: Individual parking locations track capacity, reservations, occupancy, pricing rules, and e-scooter availability, and manage vehicle park/unpark events.
Shared mobility parking: TNC and shared autonomous vehicles follow operator-level parking strategies when idle, choosing whether to park, where to park, or to remain idle in place.
Freight parking: Heavy-duty and light-duty freight vehicles use loading docks, loading zones, and curbside parking at pickup and delivery stops, with parking planned during tour optimization and executed during simulation.
Result logging: Parking occupancy, parking records, and parking choice records are periodically written to the output databases for post-simulation analysis.
The figure below provides a high-level overview of how the parking components interact across the three vehicle categories (personal, shared mobility, and freight) and how they flow into a shared parking facility agent and output databases.
Scenario Parameters#
The following scenario parameters control the en-route parking search behavior:
Parameter |
Description |
|---|---|
Simulate parking |
Master toggle for the entire parking simulation module |
Number of choices when cruising |
Number of nearby parking candidates to evaluate during cruising search |
Cruising parking distance threshold |
Maximum distance (meters) from the current link to consider parking |
Exact parking search |
Whether to use exact network distances (vs. Euclidean approximation) |
Garage distance threshold |
Maximum walking distance to consider for garage parking |
On-street parking cost |
Default hourly cost used for on-street parking when no facility data exists |
Note
Parking simulation is enabled via a scenario flag. When disabled, vehicles arrive at their destination without explicitly modeling the parking process.
The remaining sub-sections describe each of the parking components in detail.