Push a Button, Get a Bus (Part One)
Meet Alex and Menno from Routable AI. They're building a ride-sharing technology business in a pandemic.
A group of MIT students and professors published a startling paper in 2017. It showed that all of Manhattan’s taxi demand could be serviced by 80% fewer vehicles if passengers shared rides using their algorithm in real time.
Passengers sharing rides waited to be picked up for, at most, three minutes and were detoured by, at most, six minutes. This meant that a much smaller fleet of high capacity shared vehicles could potentially outperform a much bigger fleet of taxis.
Alex Wallar was one of the paper’s co-authors. He was a PhD student in the Computer Science and Artificial Intelligence Laboratory at MIT.
In April 2019 Alex co-founded Routable AI with Co-founder and CEO Menno van der Zee. Menno has a master’s degree (2018) in Systems and Control Engineering (Cum Laude) from the Delft University of Technology.
We met in May 2019 as part of an advisory role after being introduced to them by Amplify Partners VC, Sarah Catanzaro.
Dial-a-ride in real time
Routable AI is a technology company spun out of four years of work at MIT. Its core value is in a unique algorithm which approximates a solution to the dial-a-ride problem for high capacity vehicles in real time. The goal of this problem is to optimally allocate a fleet of vehicles to rider demand in a city.
Ride-sharing changed how many people travel within cities. The concept of e-hailing revolutionized urban mobility. Many people feel transportation is more reliable, safe and affordable as a result.
Ride-sharing was initially focussed on solo licensed black car rides. Push a button, Get a ride. It then spread to people sharing rides in their own cars, to products like uberPOOL, to micromobility, and now to transit. Given it started by providing an alternative to taxis (or a lack of taxis), pressing a button and getting a car felt great. Over time, people started to expect instant arrival and were disappointed when it didn’t happen.
This history and these rider expectations created algorithms that optimize for the arrival time for each individual rider and individual driver at the time of each request rather than the average efficiency of the system as a whole. As a result, many ride-sharing firms assign the closest driver to rider requests as each request comes in.
By waiting from a few seconds to a few minutes, sometimes a much more optimal schedule for all riders and drivers at that time will be found. This becomes more and more true with increased pooling. Thus when you think about routing for a bus or shuttle system at far lower prices, we are likely far better off grouping individual riders and drivers into batches for the benefit of the whole system. However, the technology to do this in real time is far from trivial.
Push a button, get a bus
Routable AI’s mission is to make mass transit feel like teleportation. Push a button, get a bus. Many governments around the world are trialling on-demand shuttles and buses but many lack the technology to make it work.
The list of failed shuttle startups is long. The efficiency you get from sharing is rarely enough to justify the discount you need to offer to get passengers to share.
The dial-a-ride problem for many passengers is difficult to approximate in real time. Routable AI’s algorithm can optimize all routes in a city to create driver schedules in less than a second.
This optimization allocates pick-ups and drop-offs as well as rebalances drivers based on past demand. It doesn’t make sense for every transportation provider to build this technology themselves and many of them lack the technical firepower to do so.
Routable AI will work with transit authorities, fleet operators, and ride-sharing companies globally to get people into higher capacity vehicles in the most convenient and efficient manner possible.
Pandemic speeds change
Shared rides are likely not going away long term but approaching the problem intelligently is critical as we go back to some new version of normal. As an example of this, in March and April, Routable AI worked with the Boston Center for the Homeless to use its technology to route homeless patients to the right bed in Boston based on their symptoms.
Routable AI can also help with things like capacity limitation in buses or the creation of on-demand bus services to replace canceled subway lines. They’re currently working with transit agencies to help them estimate the impact of capacity limitation on fleet performance using a simulation tool.
We’re living in a time of significant uncertainty (see latest Howard Marks’s Letter on Uncertainty). This brings opportunity. Many things will adjust to a new normal or completely change for some time. Those who anticipte these changes and act quickly will start great companies that help the rest of us adjust.
I believe Alex and Menno are among these wizards of change. That’s why I have recently left Canoo and joined them full time as COO at Routable AI.
I’d love to hear if you have any thoughts or questions about us!
A framework for career decision making
Tomorrow, I’ll share with you a personal framework for career decisions like the one to join Routable AI.
Much love,
Coxy
Jamescox00@gmail.com | LinkedIn | Twitter | Routable AI