Number of Trains Running on Japan’s Tokaido Shinkansen Bullet Train Line Hits Highest Ever at 383 Per Day on Average

By Hayaki Takeda / Yomiuri Shimbun Staff Writer

16:37 JST, September 8, 2025

The Tokaido Shinkansen bullet train line between Tokyo and Shin-Osaka is said to be “the world’s most tightly scheduled,” with fiscal 2024 seeing the highest number of trains running on the line ever, at 383 per day on average.

While the number of bullet trains available for use is limited, the frequency at which trains are run has increased. This is possible because of a new train operation control system that Central Japan Railway Co. (JR Tokai) introduced in spring last year.

Handwriting years

Many people think that Tokaido Shinkansen trains run at the same times and in the same places every day. However, this is not true. The operation diagrams for the train line change frequently based on factors such as season, day of the week and one-day events like live concerts by popular artists.

When the Tokaido Shinkansen Line opened in 1964, 60 trains ran each day. The number has continued to increase since then.

In fiscal 1989, the figure reached 252 per day on average, surpassing the 250 mark for the first time. The number rose to more than 300 in fiscal 2005 and to 350 a day on average in fiscal 2014.

Although the number temporarily decreased during the coronavirus pandemic, it has been rising again.

The operation schedules are completed by drawing many lines to indicate the movements of the Shinkansen trains on diagrams displaying departure times and station names. Then, JR Tokai officials analyze which trains among a total of about 170 will be allocated to each service.

By increasing factors such as the number of round trips a train makes, the trains can be operated more efficiently and the number of train services can be increased.

In the past, the allocation process was typically done using rulers to manually draw lines to connect existing lines on the diagrams.

An incredible amount of work was necessary to connect all the lines by hand while considering the various conditions. During busy times of year, such as the Bon Festival summer holiday, sometimes it took 20 hours to complete an operation diagram for a single day.

New system saves time, labor

JR Tokai needed to improve work efficiency, and the company began systematizing operation plans in 2017. In the spring of 2024, seven years after the start of its development, JR Tokai began using new system.

Staff members input into the system preset conditions such as “diagrams made daily” or “positions of the day’s first trains.” Within no more than three minutes, the message on the computer screen changes from “searching for solution” to “searching completed.”

Then, the computer screen displays an operation plan for a specific day in which the lines of the diagram are connected to allow for the same train cars to be used.

During busy periods such as Golden Week, the Bon Festival and the year-end of the New Year period, workers put the finishing touches on the plans by adding detailed revisions, which takes about two hours.

On weekdays, when there are often fewer trains running, the Shinkansen trains are sometimes operated in line with plans made by the computer system only after a human check.

The new system has been able to significantly reduce the staff’s working time and workload.

Record 483 trains per day

For Aug. 9 last year, JR Tokai made a plan to run a record 483 trains on the Tokaido Shinkansen Line. The company predicted that trains would be crowded with passengers going to their hometowns for the Bon Festival holiday and businesspeople would ride them because it was a weekday.

Such a high number was not anticipated when the system was developed, and JR Tokai officials said the scale would have been difficult to realize by the deadline if humans did the work.

However, on Aug. 8, a magnitude 7.1 earthquake occurred in the Hyuganada Sea off Miyazaki Prefecture. In addition, a Nankai Trough Earthquake Extra Information for megaquake attention was issued.

Under company rules, JR Tokai decreased Shinkansen train speeds on some track sections. On Aug. 9, delays in the schedule because of speed reductions continued until the day’s last train, but all of the scheduled trains ran.

“In addition to the strong efforts undertaken by station workers, [running all of the transit] was possible largely thanks to the fact that we were able to precisely draw up an operation plan using the system,” said Kazuhiko Hosokawa, 46, the JR Tokai official who was in charge of the day’s train operation control.

For the duration of the 2025 Osaka-Kansai Expo, which opened in April, JR Tokai has increased the number of train services of the Nozomi series of Shinkansen trains. Mainly on weekends when trains are predicted to be crowded, JR Tokai increased the number of Nozomi trains that depart stations in the Tokyo metropolitan area from 6 a.m. to 6:59 a.m to 14.

Demand for the Tokaido Shinkansen Line is predicted to continue to rise.

“I want to respond to the number of train services that will increase year by year with the system and make operation plans which are convenient for passengers,” said Masahide Koizumi, 38, a JR Tokai employee who was engaged in the development of the operation control system.

AI to be used for train control

Railway companies need to work quickly if trains are delayed, and schedules must be altered due to accidents or other troubles.

NEC Corp. developed an artificial intelligence system that can generate the most suitable diagrams in a short time after such issues.

The company has repeatedly conducted tests to raise the accuracy of the AI system, aiming to put it into practical use as soon as possible.

To minimize the negative effects on diagrams of successive trains, revising the diagrams requires both speed and accuracy. However, in recent years, an increasing number of railway companies have begun jointly operate connected services, in which one company’s trains directly run on another firm’s rail lines. As a result, the types of trains, the number running and other operational factors have become complicated.

In addition, in the workplaces in charge of making diagrams, the number of veteran employees with extensive knowledge of the work is decreasing. Therefore, a labor shortage has become a serious problem.

When developing the AI system, NEC adopted “enforced learning,” a method that is also used for shogi plays between AI programs to increase their performance.

While obtaining cooperation from railway company officials, NEC made the AI system respond to records of various cases in the past for normalizing train diagrams. Through the process, NEC had the AI system learn how to output optimal solutions.

Kyushu Railway Co. (JR Kyushu) and Hitachi, Ltd. are jointly working to utilize AI technologies for the revision of diagrams in the event of any issues.

“While reducing the burden on employees, we want to shorten the time until the train schedule will be established and improve the quality of our transportation services,” said a JR Kyushu official.