Quality of service metrics for risk-aware car sharing operations with substitution

This is the latest version of this item.

Official URL: https://dx.doi.org/10.1016/j.omega.2026.103569

In today's competitive car sharing industry, where demand for flexible transportation is rising, optimizing operations while ensuring high service quality is essential for efficiency, profitability, and competitiveness. This paper considers a reservation-based car sharing company that offers round-trip and one-way services using a mixed fleet from different market segments to address diverse customer preferences. The problem involves determining service station locations, allocating vehicle types to stations, and operational planning over a finite horizon to maximize expected profit in a medium- to long-term planning setting where decisions are made prior to demand realization. Demand uncertainty across vehicle types is modeled using a finite set of scenarios within a two-stage stochastic program. We incorporate substitution options, enabling the company to offer alternative vehicle types when the preferred ones are unavailable, and model customer preferences via a utility-based framework capturing satisfaction levels and acceptance probabilities of substitution. We develop quality of service metrics for risk-aware car sharing operations, leveraging conditional value-at-risk to capture utility variability and enforce minimum satisfaction across customer groups and service regions. Beyond profitability maximization, the proposed framework enables the systematic evaluation of strategic trade-offs between budget levels, service targets, and risk preferences, and provides insights into how these configurations affect customer satisfaction and substitution patterns. To solve this complex problem, we propose a decomposition-based branch-and-cut algorithm. Our computational study shows the effectiveness of the proposed approach in optimizing profit and service quality, solving up to 5.5 times more instances than the commercial solver within the same time limit.