Dynamic Ridesharing Problem with Meeting Point Extension
Ridesharing operators provide a service with which passengers with a similar direction and period of travel can share a vehicle. As urban congestion continues to worsen and emissions increase, it is controversial whether ridesharing services can contribute to improving the traffic situation in cities. Studies of the static ridesharing problem and practical implementations by Uber and MOIA show that ridesharing can be made more traffic-friendly by introducing meeting points. Meeting points are predetermined locations where passengers get on or off a ridesharing vehicle. Passengers are willing to walk to these locations and thus cover part of the journey on foot.
This thesis examines the effects of an extension of the dynamic ridesharing problem by meeting points. Experiments on a large-scale real-world data set show that also in the dynamic case, the introduction of meeting points results in an increase in the number of assigned trips and a reduction in the traffic load. Thereby a dependency of the results on the input parameters fleet size, customer waiting time, customer travel time and the selection of the meeting points is determined. Although meeting points contribute to a reduction in traffic caused by ridesharing services, the ridesharing services does not seem to be a recommendable alternative to private cars in the studies carried out.