Platform-induced time-space trade-offs in ride-hailing: Multi-homing as a response to operational constraints

This study examines how ride-hailing drivers adjust their time-use and spatial behavior under platform-induced constraints, with a focus on multi-homing—the practice of operating across multiple ride-hailing platforms. Drawing on a city-scale, driver-identified dataset from Suzhou, China, we propose a data-driven framework to identify multi-homing behavior and quantify its impacts using four operational metrics: working hours, travel distance, revenue, and order interval. A common assumption is that full-time multi-homing drivers earn more and work longer than single-platform drivers. However, our results show that this assumption does not hold in the Suzhou market. Instead, multi-homing appears to serve as a behavioral adaptation to regulatory and algorithmic restrictions—allowing drivers to bypass platform-imposed work-hour caps and optimize engagement with temporal demand fluctuations. Using clustering to separate full-time and part-time drivers, and applying Geographically Weighted Random Forest (GWRF) modeling, we further find that multi-platform activity is not spatially concentrated in low-demand or remote areas. These findings reveal that multi-homing is less about spatial expansion and more about temporal strategy and coping with institutional uncertainty. The study contributes to understanding time-space adaptation in digitally mediated mobility, especially amid evolving platform governance. It also underscores the need for time-use models and transport policy to account for the real-time flexibility and constraint navigation strategies employed by gig workers in fragmented digital environments.