Over the last decade there has been a surge of short-term rental schemes and shared-use mobility concepts. As defined in by Dr. Shaheen, shared-use mobility is “an innovative transportation solution that enables users to have short-term access to transportation modes on an as-needed basis.”. Under this broad concept of shared-use mobility we consider any means of transportation that serves the above purpose, including carsharing, peer-to-peer carsharing, bikesharing, ridesharing, ridesourcing, among others.

Several companies are taking different approaches to provide shared-use mobility concepts. For example, ride-hailing companies such as Uber and Lyft started offering on-demand mobility, but quickly launched Uperpool and Lyft Line to let their customers share rides. According to Lyft CEO, Lyft Line became within 6 months the most popular service the company is offering in San Francisco. Others, such as Leap or Bridj offer an alternative to transit by providing a flexible network of private bus lines. ZipCar provides short-term rentals, Carma, former Avego, lets users log their commute trips so users can find matches, and PiggyBack automatically creates carpools on a per trip basis.

Numerous benefits have been reported of such a systems, including: car-ownership and vehicle usage reduction, increase network connectivity and encouraging multi-modality. However, its long-term effects are yet to be understood. There is an interesting discussion in this mobilityLab post about the effects of shared-use mobility.

In addition, most, if not all the approaches presented above do not fully integrate with the existing transit network, instead it is left to the users to decide whether or not is more efficient to transfer to transit along their trip. Furthermore, most of the companies focus their efforts on urban areas with rather large population densities, leaving out residential areas where it may be more costly or difficult to provide the service.

In line of this observations, Car2work is a new shared-use mobility concept that its main goal is to connect commuters with workplaces, guaranteeing a trip home, and connecting with the existing transit network, efficiently tackling the last mile problem of transit.

In such proposed system, the vehicles do not belong to the commuters, instead an operators owns and assign the vehicles to the commuters. While the vehicles are idling at transit stations or workplaces, they can be rented as a traditional short-term car rental service.


I used Google Maps to check the travel times from University of California, Irvine to Union Station, Los Angeles, both using a private auto and transit.

As it can be observed from the images below:

  1. It is 50 min faster to drive than taking transit (UCI-LA)
  2. Using transit it can take more than 1hr to travel 8 miles (UCI-Transportation Center)
  3. Almost all time savings occur on connecting UCI with Transit Station




How it works

  1. Commuters announce their (multiple) trips in advanced (1 day) and thir preferences.
  2. Vehicles do not belong to the commuters, instead there is an operator that owns them.
  3. Commuters do not have a vehicle assigned and:
    • May or may not take transit link.
    • Can specify transfer preferences.
    • Can share rides with commuters from other companies.
  4.  An optimization problem is solved to match vehicles to commuters:
    • If a commuter is matched, the system guarantees the return home trip.
    • If a commuter is not matched, it is assumed that it has other means to make the trip.
  5. While vehicles are idling at transit stations or workplaces, they can be rented as a short-term car rental service (ZipCar) to anyone enrolled in the program.

 An example

In this case we have an example with 3 workplaces, 5 commuters and 3 vehicles.


Solid arrows represent the morning commute and dashed arrows the evening commute.

Simulation Results

The table below summarizes some of the preliminary results that I have and points to an interactive visualization.
The first case shows the solution when trying to maximize the commuters matched and minimize total vehicle travel time. The second,  when maximizing commuters matched and minimizing total users travel time.

This second option provides a solution with less lingering for commuters but slightly higher vehicle travel times.

# Vehicles# W-places# UsersCommentsLink
4310Minimize vehicle travel timeGo
4310Minimize user travel timeGo