Robust heterogeneous fleet management for dual-sourcing on-demand ride services under demand uncertainty

Motivated by the emerging dual-sourcing strategy with a blended workforce in on-demand ride service markets, this paper investigates the impact of this strategy on ride services and sizes a heterogeneous fleet operated by both contractual and freelance drivers. We present an integrated heterogeneous fleet management model that incorporates strategic fleet sizing and tactical platform operations, addressing demand uncertainty through an adaptive passenger–vehicle matching rule that outperforms conventional static methods. Employing a two-stage robust optimization framework, the second stage is reformulated into equivalent mathematical programming formulations, solvable using state-of-the-art commercial solvers. A column-and-constraint generation based exact algorithm is developed to solve the two-stage model by leveraging its structural properties. Extensive experiments based on a real-world ride-hailing service demonstrate significant benefits of the dual-sourcing strategy in enhancing service performance. • Heterogeneous fleet management for dual-sourcing on-demand ride services. • Integrated model that incorporates strategic fleet sizing and tactical operations. • Two-stage robust optimization model based on an adaptive matching rule. • Column-and-constraint generation algorithm for exact solutions.