22 November 2023 in Journal of Cleaner Production
19 May 2026 in IEEE Transactions on Engineering Management
This study proposes a digital twin (DT) model for drone logistics systems. While DT models typically rely on real-time data from physical assets, identifying and structuring such parameters for drone logistics remains underdeveloped in current literature. Existing studies primarily focus on routing and fleet optimization, offering limited guidance on the systematic design of DT architectures for drone operations. To address this gap, this work conducts an evidence-based, structured review of 116 drone logistics models to extract and consolidate key operational parameters, resulting in the development of a Drone Logistics Parameter Framework (DLPF). Building on this foundation, the study introduces the Drone Logistics Digital Twin (DLDT) architecture and model. The DLDT is a five-dimensional system consisting of a physical entity, a virtual entity, a service interface, a data layer, and connection layers. The DLDT enables essential capabilities, including real-time monitoring, predictive maintenance, route optimization, last-mile delivery management, and scheduling. To validate the DLDT architecture and model, a route-optimization case demonstration within the DT environment illustrates their internal coherence and practical utility. Hence, this study contributes to a structured approach for designing DT models in drone logistics, addressing a notable gap in the literature and offering actionable insights for advancing intelligent, data-driven logistics operations.
14 August 2025 in Transportation Research Record: Journal of the Transportation Research Board
Minimizing fuel consumption and carbon emissions through vehicle routing is a critical objective for logistics companies, especially in light of stringent greenhouse gas emission policies enforced by various countries. To address this challenge, numerous studies have proposed robust vehicle routing problem models. Many of these focus on heterogeneous vehicle routing with time window constraints, aiming to reduce fuel consumption and emission costs. However, most research treats delays in customer time windows merely as penalty costs, giving limited attention to the impact of these delays on CO 2 emissions and their associated costs. Such delays—often caused by traffic congestion and rerouting—are significant contributors to excess CO 2 emissions. This paper addresses this overlooked aspect by proposing a model that accounts for the additional fuel consumption and CO 2 emissions resulting from delayed deliveries. It explores uncertainty in delivery schedules by analyzing three critical traffic scenarios: (1) vehicles idling in standstill traffic, (2) vehicles moving in slow traffic, and (3) vehicles being rerouted because of traffic uncertainties—all of which exacerbate CO 2 emissions from late deliveries. We propose a heterogeneous vehicle routing problem with carbon penalty, which optimizes routing to minimize CO 2 emissions for both on-time and delayed deliveries. A mathematical model and a robust genetic algorithm are developed, with their effectiveness validated through practical scenarios. The results offer valuable insights for logistics providers seeking to enhance both sustainability and operational efficiency.