We propose a robust model to study the multi-period assignment problem in ridesharing in which both demand and travel time are uncertain. We consider finite pickup-and-delivery locations and the demand at each location can be uncertain. Although we could obtain historical data, but it is hard to get full distributional information of demand. Moreover, we also consider uncertain travel time between origin and destination, since in real world, traffic condition is affected by many factors such as time, weather and others.

We start from the assignment problem with only uncertain demand, which could be formulated as a stochastic optimization model. After introducing factor-based demand model, we then apply static and linear assignment policy to reformulate the original problem to a t...[ Read more ]

We propose a robust model to study the multi-period assignment problem in ridesharing in which both demand and travel time are uncertain. We consider finite pickup-and-delivery locations and the demand at each location can be uncertain. Although we could obtain historical data, but it is hard to get full distributional information of demand. Moreover, we also consider uncertain travel time between origin and destination, since in real world, traffic condition is affected by many factors such as time, weather and others.

We start from the assignment problem with only uncertain demand, which could be formulated as a stochastic optimization model. After introducing factor-based demand model, we then apply static and linear assignment policy to reformulate the original problem to a tractable one. In a more general model, which considers both uncertain demand and travel time, we conduct a simple analysis where travel time varies in two possible values. In the computational study, we use real data from DiDi GAIA Open Dataset to evaluate the performance of our model. Furthermore, the results show that our model could achieve or exceed the performance of some heuristics derived from dynamic programming.