Scientists have created a system that has solved one of the most complex problems in public transport management. For the first time, the Russian AI model was able to simultaneously take into account three factors: traffic schedules, vehicle deployment, and driver shifts. Previously, such calculations were considered too complex even for modern systems. The development was reported by the Russian Ministry of Transport.
The model was created by the scientific team of "Opturana" – a Russian platform for transport resource planning. The system is designed to help calculate public transport operations not in separate stages, but as a single task.
Typically, urban transport planning proceeds sequentially. First, a traffic schedule is drawn up, then based on it, which buses will be deployed on routes, and after that, driver work schedules are formed. This approach can lead to an excess of vehicles on the line, overtime, and inconvenient shifts.
The Russian AI model takes into account all key parameters at once: traffic intervals, transport capacity, driver work and rest regimes, and other restrictions. As a result, the system looks for not separate solutions for each stage, but a single overall option for the entire transport complex.
Previously, the complexity of calculations prevented combining all three tasks simultaneously. According to the Ministry of Transport, leading global developers usually solve a maximum of two out of three tasks together. The "Opturana" team was able to overcome this limitation with the help of artificial intelligence.
The system is capable of creating and evaluating millions of possible traffic and crew operation options, simultaneously checking restrictions in all directions. As a result, a single coordinated plan is formed, rather than several separate schemes that have to be manually adjusted to each other.
The new technology has already undergone practical testing. For two weeks, buses on nine routes of the "Andropova" site of GUP "Mosgortrans" operated according to the schedule built by the system.
According to the test results, inter-trip downtime was reduced by 34%, drivers' night work was decreased, and two buses were freed up and redirected to reinforce other routes.
In addition, calculations showed a reduction in the payroll fund by 5.2%, which amounts to 13.4 million rubles per year. The new technology can be scaled to other regions of Russia.