Trading Flexibility for Adoption: Dynamic vs Static Walking in Ridesharing

Host(s): Associate Professor John Turner

Area: Operations and Decision Technologies

Speaker(s): Julia Yan, Assistant Professor of Operations and Logistics

University: University of British Columbia, Sauder School of Business

On-demand ridesharing aims to fulfill riders' transportation needs whenever and wherever they want. Although this service level is appealing for riders, overall system efficiency can improve substantially if riders are willing to be flexible. Here, we explore riders' flexibility in space via walking to more accessible pickup locations. Ridesharing platforms have traditionally implemented dynamic walking, which jointly optimizes rider-driver assignment with rider pickup locations. We propose the new paradigm of static walking, which communicates a predetermined pickup location to the rider, and then optimizes rider-driver assignment. On its surface, dynamic walking appears to be the gold standard; the flexibility of optimization as compared to the restriction of a predetermined pickup location makes the viability of static walking far from a foregone conclusion. However, a major drawback of dynamic walking is that riders are deterred by the associated uncertainty. Static walking aims to mitigate this uncertainty, as a semi-flexible approach that achieves the value of flexibility without burdening riders. We study characteristics of networks on which static walking can be viable and propose algorithms to generate fixed pickup locations. In simulations on Lyft data in Manhattan, we found that static walking achieved up to 94-95% of the value of dynamic walking; static walking can therefore overcome dynamic walking's advantages with just a modest relative increase in rider adoption of as low as 5-6%. We support our simulations with empirical evidence that static walking has substantial value using data from hundreds of thousands of Lyft rides.