Maximizing public transportation capacity is an important way to ease urban traffic congestion and improve the travel experience of city residents.

  • Bus schedules change and it may take a while to adjust schedules because this process relies on individual experience.

    Bus schedules change and it may take a while to adjust schedules because this process relies on individual experience.

  • Bus scheduling is not based on a comprehensive analysis of the demand and supply during different hours of the day. Sometimes, this means that supply can't match demand and passengers have to wait longer for their buses during peak hours.

    Bus scheduling is not based on a comprehensive analysis of the demand and supply during different hours of the day. Sometimes, this means that supply can't match demand and passengers have to wait longer for their buses during peak hours.

  • Bus scheduling is an optimization problem that involves numerous factors. Manual bus scheduling is both challenging and time-consuming.

    Bus scheduling is an optimization problem that involves numerous factors. Manual bus scheduling is both challenging and time-consuming.

Shenzhen Bus Group is a state-owned public transportation operator. SZBG provides urban public, tourist, and long-distance passenger transportation, along with bus leasing, advertising media, station construction, property management, new energy charging, and more. It is the first executive member of the International Association of Public Transport (UITP) committee and a rotating chair of its Taxi and Ride-Hailing Committee from China.

Three core capabilities of Huawei Cloud's Intelligent City Hub scheduling engine

Intelligent bus scheduling powered by operations optimization algorithms: Quickly generate initial bus schedule plans for over 6000 buses, 300 bus routes, and considering 20 constraints.

A real-time scheduling algorithm that adapts to dynamic conditions: Using the initial bus schedule and real-time information, a real-time scheduling algorithm automatically updates bus arrival times, making sure that the operations center, drivers, and passengers can all obtain accurate arrival predictions.

A multi-dimensional spatiotemporal model: A Gradient Boosting Decision Tree Algorithm (GBDT) model mines aggregated operational data for multi-dimensional spatiotemporal features, assisting the real-time scheduling engine in accurately predicting bus arrival time.

Customer Benefits

  • Bus arrival time predictions are accurate to less than 2 minutes, buses are 90% punctual, and passengers spend 10% less time on their journeys.

    Bus arrival time predictions are accurate to less than 2 minutes, buses are 90% punctual, and passengers spend 10% less time on their journeys.

  • A bus scheduling plan can be customized daily and generated in just 3 minutes.

    A bus scheduling plan can be customized daily and generated in just 3 minutes.

"Shenzhen Bus is working with Huawei Cloud to convert computing power into transport capacity. We will continue to transform public transportation using digital technology and explore new business models together."

Liao HanqiuDeputy Manager of Intelligentilization Dept, Shenzhen Bus Group


Liao Hanqiu
Deputy Manager of Intelligentilization Dept, Shenzhen Bus Group

"Shenzhen Bus is working with Huawei Cloud to convert computing power into transport capacity. We will continue to transform public transportation using digital technology and explore new business models together."