Uncertainties in shared mobility optimization problems: Survey and perspective

Shared mobility systems encompass a wide range of decisions, from real-time operations to strategic long-term planning, all influenced by inherent uncertainties. This review categorizes these uncertainties into four primary types: demand, supply, behavioral, and operational environment, and examines their unique impacts on strategic, tactical, and operational decision-making processes. We review traditional and emerging optimization methodologies, including stochastic programming, robust optimization, data-driven approaches, and reinforcement learning. By linking these methodologies to specific uncertainty types and decision-making levels, this review provides insights into their applications, strengths, and limitations. Finally, we identify research gaps and highlights opportunities for advancing the optimization of shared mobility systems under uncertainty. • Identifies and categorizes four primary types of uncertainty in shared mobility systems: demand, supply, behavioral, and operational environment uncertainty. • Systematically analyzes the impact of these uncertainties on strategic, tactical, and operational decision-making levels across multiple shared mobility services. • Reviews and synthesizes methodologies, linking them explicitly to specific uncertainty types. • Highlights the unique challenges and suitability of each optimization method under various uncertainty conditions within shared mobility contexts. • Provides comprehensive insights into the current gaps and future research directions.