Optimization over multiple objectives
Mobi allows users to search for routing directions that through almost all modes of ground transportation options, including driving, cycling, transit, taxi, bike sharing, ridesharing and car sharing. Through combining the different public and private transportation networks, Mobi is able to provide more and better options for the users than single-modal routing services in many scenarios.
Through the integration of a rich preference and constraint models, Mobi allows the users to specify constraints, such as trip length, cost and walking time. In order to provide users with more desired solutions to choose from, routing results are optimized along the pareto front of two or more objectives, such cost, duration and walking times. No matter they are looking for the fastest route to catch an important meeting, or the most economical ways to get to Fenway Park for a baseball game, our intermodal-routing algorithms can find the mode combinations that best fit their needs.
Most of our trips in city are intermodal, even though we often do not realize that. When we are driving to a restaurant we need to park nearby then walk to the destination. When we are going to a field trip it is often faster to take transit. Intermodal because of the following advantages:
Connectivity: no single transportation network or service provider can cover all corners of the city. Walking and biking only make sense when our trips are short. Transit is only a viable option when our origin/destination are near a stop, and most of their services terminate at night. Driving seem to be a good 24/7 option, but only if traffic is good and parking is available at the destination. Intermodal routing knit the different public and private services into a much larger and dense network, making our city more accessible throughout the day.
Efficiency: similarly, no single transportation network or service provider can claim that they are the fastest option anytime and anywhere. At night and in the suburbs, driving and taxi are clearly the fastest options. While during rush hours, rapid transit is the winner in speed, especially in downtown areas with lots of traffic. This is why many people choose to park their cars at the terminal subway stations, then take the trains to their downtown office. Intermodal routing allows us to utilize the most efficient portion of different transportation modes, and achieve higher efficiency than any single mode options.
Convenience: intermodal routing also improves the convenience of our trips, either by reducing the distance of walking, or the amount of waiting times. For example, if no parking is available at the destination, we may choose to park at a garage close to a bikeshare station, and take a short ride on bicycle. Or if the queue for taxi is extremely long at the airport, we can first take the bus to a region with better availability, then get off and switch to taxi. Neither of which can be achieved if we locked ourselves in one mode.
Cost-effective: driving is often the top choice for American car owners, but the traffic and parking cost/toll can easily make it the slowest or most expensive option (sometimes it gets both title). Intermodal routing enables the combination of driving your personal cars with other mode of transportation, in order to avoid the congested segments and lower parking cost. For example, if you are going to Fenway Park to watch baseball, you may park at a distance to avoid the inflated rate due to the game. It even saves you money and walking distance by parking in Cambridge then Uber to the stadium.
However, given that intermodal routing is such an important problem, it is surprising that there has not been much research (at least published) on it. Unlike transit routing, which is also a challenging problem but has been studied extensively by scientists at universities and research labs (see Google and MSR’s publications), people have not published any effective approach on intermodal routing, except for acknowledging its difficulty and a few potential directions.