The primary goal of the proposed study is to develop tools that can be used to enhance the performance of public Bicycle Sharing Systems (BSS), in particular the Capital Bikeshare. Capital Bikeshare is one of the oldest and largest BSS operating in Washington D.C., Virginia, and Maryland. In this context, the objectives of the study are three-fold:

Objective (1): Develop real-time system state prediction models of BSS. These models will be able to predict the number of departures of customers for retrieving bikes and number of arrivals of customers for returning bikes at each station by time-of-day.

Objective (2): Develop heuristic algorithms for managing the expected demand patterns. This includes development of quick heuristic algorithms that can identify optimal rebalancing schedules as demand evolves in real-time

Objective (3): Integrate the predictive models into a GIS toolkit with an easy-to-use Graphical User Interface (GUI) that visually depicts the future demand patterns (for BSS customers) and associated optimal rebalancing routes and schedules (for BSS operators).

The envisioned models to accomplish the study objectives encompass several methodological innovations both in the statistical and optimization domains. To ensure that the transfer of these models to practice is smooth and efficient, these models must be packaged into an easy-to-use tool that BSS agencies can use for their operational planning purposes. To this end, this research will develop a GIS toolkit that will generate color-coded intensity maps that depict the current and expected demand patterns at all stations along with a set of optimal routes and schedules for rebalancing. The potential customer base for this tool include (a) the BSS customers and operators as well as local transit agencies who intend to serve the public travel needs better and enhance their respective ridership levels, and (b) public health officials who are interested in identifying the best practices to promote active transportation. During the course of this project, the research team will continuously interact with the Capital Bikeshare officials to construct plausible future scenarios concerning station locations, station capacities, and rebalancing fleet size instead of considering purely hypothetical scenarios.

The study findings will help improve the performance of Capital Bikeshare by identifying key temporal and spatial demand patterns and associated optimal rebalancing strategies under a host of different policy scenarios. Also, the proposed methodological framework that ties together the demand and supply dimensions will serve as a good modeling paradigm for analyzing other BSSs. Given the ubiquitous nature of count data in the real world, the econometric models developed in this study for modeling BSS demand will see wide applicability both within and outside transportation.

http://www.tri-odu.org/sustainability–green-transportation.html

Final Report: Real-Time System Prediction & Optimal Rebalancing Strategies for Public Bike Sharing Systems (Submitted February 2018)