Memo

Abstract

In this assignment, my team and I wrote a memo that would be addressed to Rafael Portnoy, the CTO of the MTA, highlighting a change that needs to be made in our MTA systems. We highlighted the untrustworthy scheduling and frequent delays of the trains in NYC, and suggested the implementation of AI and deep learning to fix this issue so that trains run on stricter time schedules and repairs and maintenance can be managed in a more organized matter.

Memo

To:         Rafael Portnoy, CTO of the MTA

From:    Abdelrahman Ahmed, Topic Supervisor

               Nabil Omi, Research Director

               Susmita Halder, Submission Manager

Subject: Using deep learning and artificial intelligence to improve our transportation systems.

Date:      March 28, 2021

The purpose of this memo is to focus on New York’s old, inconsistent, and delayed subway systems and explain how implementing AI runs, human-approved scheduling can vastly improve the passenger experience through greater consistency and lesser delays.

Summary

The New York subway system is very antiquated, falling well behind many other metropolitan centers in the world such as Hong Kong. Compared to what New Yorkers have, the Hong Kong subway system has a far higher on-time record, at 99.9% of the time. Anyone who has ever taken the subway knows just how unreliable our current system is. To achieve that task, they utilize AI to plan out schedules of repairs, re-routes, and other activities. Utilizing AI and New York’s massive potential for computing power, the subway system can be far better than it is by making it more standardized and consistent.

Discussion

In New York City out of all the people who use public transit commute to work, 39% use the subway. But the most common problem in the subway system is delays. The major incidents causing subway delays are caused by tracks, signals, medical, station and structure, subway cars, and others. Unlike many other countries in New York, the metro lines share the track with each other. Only the subway lines: 1,6,7 and L have dedicated lines, where 2,3,4, and 5 only share tracks with each other, and the remaining 13 subway lines all share tracks at various points. Therefore, delays on one line propagate to other lines. The subway system also has been used for more than 100 years, as do its signals systems. The signal problem is one of the most top reasons for the delays. According to The New York Times article, only two of the subway’s 22 routes the L and 7 have upgraded signal systems and have the highest on-time rates. The L-train has a 93 percent on-time rate whereas the A train has an on-time rate of 65 percent. The most important system to run the whole city New York subway system has not the best on-time record comparison with the fact that Hong Kong has one of the best subway systems with its 99.9 percent on-time record. The reason for Hong Kong’s best subway system is artificial intelligence (AI), which schedules and manages from grinding rough rails smoothly and replacing tracks to check for damages. 

Recommendation

One of the main cities today that has implemented AI technology into its subway systems is Hong Kong. They use this system to plan repairs and works to be done by the engineers that work for the MTR, which is the company that runs the subway system in Hong Kong. Instead of having people from multiple areas meet to discuss plans for repairs and maintenance, the AI does all the work, allowing for a much more efficient system. It has one of the best on-time records out of any other major subway system in the world, a 99.9%. Since New York City’s subway system is one of the very few 24 hour systems in the world, it has the least time to run repairs and maintenance. Having artificial technology take control of this, planning out the most efficient way to perform these repairs and maintenance would have a major long-term benefit in terms of efficiency, timeliness, and even cost-effectiveness. New York City is very dependent on its subway system, which is quite unreliable. Considering that 56% of the population uses the public transport system, it needs to be more efficient and get people to their jobs on time, since most of these people are commuting to and from work. The use of artificial technology is the best way to achieve this and will benefit New York City in all aspects.

One of the best ways to build a system that can efficiently do this while also helping fund some public universities at the same time would be to get graduate students to help assist in this endeavor. Initiating a program for people in the fields of data science, computer science, several engineering fields, and also known professionals in the field of AI to help them. This allows cheap collaborative development of a system that would have a great real-world tangible effect and also helps students gather more information and experience with AI. Not only would this system help the people working on it by giving them experience with professionals, but it would also allow the city to (compared to contractors) cheaply develop a scheduling system that would be approved by an experienced professional prior to development, just like what Hong Kong already does.

We are grateful for your review of this proposal and would like to talk to you more about this initiative whenever you are available. We can invite many others to participate in the discussion of this very important topic and take initial steps.

Work cited (APA):

Public transit facts. (n.d.). Retrieved March 28, 2021, from https://nytransit.org/resources/public-transit-facts

Skip to toolbar