Improve service quality for bike sharing

Bike sharing has become commonplace in many places, and you've certainly seen them in various Swiss cities: Electric bikes or e-scooters in bright colors that can be used at train stations and other public places - usually via app - to get from A to B quickly. But often enough, these vehicles are parked somewhere after use, for example on sidewalks, in building entrances or elsewhere in public space. There they often become a nuisance for other road users or for residents. And they have to be collected again by the city bike providers at great expense.

Users' expectations of bike sharing are diverse: They want a vehicle that is quickly available and functional. And they want it where there is a need for it, i.e. not only at train stations but also, for example, at concert halls, sports stadiums or restaurants, so that they can get home quickly and safely from there. A study by the Norwegian University of Technology and Natural Sciences (NTNU) in Trondheim has therefore looked at how cities and bike-sharing providers can improve the service and also traffic management.

How to shoot at a moving target

Providing bicycles or e-scooters where and when people need them is a challenge. The problem is described as dynamic because it is constantly changing, and stochastic because it changes in random ways that are often difficult to predict. Steffen Bakker, a researcher at NTNU's Department of Industrial Economics and Technology Management, explains it as follows: "Users of the bike-sharing system pick up their bikes at one location and then take them to another location. Then the state of the system changes because suddenly the bikes are not where they were originally, that's the dynamic part," he said. "On top of that, you don't know when the customers will pick up the bikes and where they will leave them. That's the stochastic part. So if you want to plan at the beginning of the day, you don't know what's going to happen." It's like shooting at a moving target, he said. In other words, what's desired is a system that can make more accurate predictions about where and when there will be increased demand for bicycles and e-scooters. Bakker and his researcher colleagues have therefore developed an optimization model that makes recommendations to bike-sharing operators on how to dispatch bikes and scooters, as well as their service vehicles. The aim is to improve the process of so-called "rebalancing", i.e. collecting and transporting bikes from one parking station to another.

Assemble the parts correctly

The Norwegian researchers conducted a pilot test in Trondheim for this purpose. "With this, we want to use existing city bike systems as a test base and increase the efficiency of the rebalancing teams by 30 % and the lifetime of the bikes by 20 % by developing new decision support tools," said Jasmina Vele, project manager at Urban Sharing, the bike-sharing company involved in the research project. "This can be achieved through better rebalancing and preventive maintenance decisions, which will lead to a big cost reduction in existing urban bike systems." The optimization model, which is still in the development phase, can be used to communicate a new plan to service vehicle riders each time they arrive at a bike station.

That's the tricky part. It's important not to be too myopic and focus only on the current state of the system, Bakker says, especially if certain stations are expected to have more demand in the next hour or so. "It's very complex because it's a big system," he says. "Maybe in an hour there will be a big demand at the station. So you already want to get some bikes there. But at the same time, there may be stations that are almost empty now and need bikes. So you have to find a compromise."

Bike sharing modeling with digital twin

Bakker and his colleagues are working with NTNU's Department of Computer Science to create a "digital twin," or computer simulation of the systems. This allows them to test different models and try different approaches without having to test them in the real world. Initial tests have shown that the model created by the group can reduce the number of problems (i.e., either too few bikes in the location where the user wants one, or too many bikes so that the user cannot park his or her bike) by 41 % compared to not rebalancing.

The team around Steffen Bakker has also been working on a component of the optimization model called the criticality score. A criticality score is basically a score assigned to different bike share parking spaces based on the number of bikes they currently contain or require. These scores are relatively easy to calculate and can be provided to riders as they travel around the city to balance the number of bikes at each station. "It's a score that tells the service rider which station they should definitely go to," Bakker said. "It allows us to offer something that, while not the best, is probably good and much better than what bike-sharing companies currently have available." Urban Sharing's Jasmina Vele, then, confirms that using these types of optimization models can help make bike sharing an important part of urban transportation. "Urban Sharing's vision for future mobility is a transportation system that is responsive and adaptive. By using data and machine learning/optimization algorithms, we can combine the best of traditional and modern transportation systems to create a resource-efficient system that responds to demand and adapts to individual user needs," Vele said.

The research paper was published in the European Journal of Operational Research. Source: Techexplore.com