NEC has built a prototype system that utilizes AI to create optimal recovery schedules in a short period of time in the event of transportation disruptions due to accidents or natural disasters.
This system utilizes AI “large-scale system optimization technology” equipped with reinforcement learning that can solve large-scale optimization problems in a realistic amount of time, realizing the creation of an optimal recovery schedule in a short time. Using a digital twin for railway operation that reproduces railway operation in a digital space, we will verify the practicality of the created timetable.
In addition to actual transport disruption cases, by generating inexperienced cases on the digital twin, AI learns how to deal with them through trial and error.
Conventional combinatorial optimization AI searches for the optimal solution with a huge amount of calculations when transportation is disrupted. On the other hand, with the optimization AI this time, by learning how to deal with it in advance, it is possible to create an efficient timetable in a few minutes in the event of a transport disruption.
This system was technically verified with a digital twin for railway operation on the Odakyu Odawara Line.