Research Project: A New Perspective on City Logistics: Concepts, Theory, and Models for Designing and Managing Logistics as a Service
Loading...
Contributors
Funders
ID
EC.00179
Authors
Yıldız, Barış
Faculty Member
Publications
Express shipments with autonomous robots and public transportation
(Elsevier, 2024-12-01) Yıldız, Barış; Salman, Fatma Sibel; Ermağan, Umut; Department of Industrial Engineering; Yes; College of Engineering
Growing urbanization, exploding e-commerce, heightened customer expectations, and the need to reduce the environmental impact of transportation ask for innovative last-mile delivery solutions. This paper explores a new express shipment model that combines public transportation with Autonomous Robots (ARs) and studies its real-time management. Under dynamic demand arrivals with short delivery time promises, we propose a rolling horizon framework and devise a machine learning-enhanced Column Generation (CG) methodology to solve the real-time AR dispatching problem. The results of our numerical experiments with real-world delivery demand data show the significant potential of the proposed system to reduce travel time, vehicle traffic, emissions, and noise. Our results also reveal the efficacy of the learning-based CG methodology, which provides almost the same quality solutions as the classical CG approach with much less computational effort.
A machine learning approach for marginal fulfillment cost estimation in last mile delivery
(Elsevier, 2025) Nalbant, Ali; Yıldız, Barış; Nalbant, Ali; Graduate School of Sciences and Engineering; Department of Industrial Engineering; Yes; GRADUATE SCHOOL OF SCIENCES AND ENGINEERING; College of Engineering
Determining marginal fulfillment costs (MFC) is crucial for effective decision-making in online grocery retail, a sector struggling with small profit margins and arduous service requirements of attended home deliveries. Paramount to improving operational efficiency, e-grocers need accurate real-time MFC estimations to optimize their service offers and prices for online customers. Traditional methods for estimating MFC are either too slow for online decision-making or inaccurate. This paper introduces a novel machine learning (ML) approach that provides fast and accurate MFC estimations with the help of carefully engineered features (predictors) that can capture complex routing dynamics. Experiments with real-world data demonstrate the superiority of the proposed approach over state-of-the-art MFC estimation methods. Our analysis of more than 2000 potential predictors, from which 20 are curated for practical applicability, reveals critical insights into the use of network-level, neighborhood-based, and node-level features in capturing complex VRP dynamics to develop ML-based approaches to address problems that arise in different transportation applications.
Crowdshipping problem with dynamic compensations and transshipments
(Pergamon-Elsevier Science Ltd, 2024) Şardağ, Ali; Yıldız, Barış; Kizil, Kerim U.; Department of Industrial Engineering; Graduate School of Sciences and Engineering; Yes; College of Engineering; GRADUATE SCHOOL OF SCIENCES AND ENGINEERING
Rapid urban growth and consequent increase in e-commerce demand make urban logistics a harder task than ever. The growing size of urban delivery operations not only entails operational challenges but also generates several negative externalities, such as increased traffic, pollution, noise, and accidents. This trend creates a pressing need for efficient delivery mechanisms that are more economical and environmentally friendly than existing systems. Crowdshipping, wherein ordinary members of the society partake in delivery operations for a small compensation, is one of the answers that cater to this need and has attracted considerable research interest recently. However, designing compensation mechanisms to prompt efficient participation from the public remains largely unexplored in the literature. In this study, we devise a dynamic compensation scheme for crowdshipping operations in a many-to-many express delivery framework, where the crowdshipper compensations are determined based on spatial and temporal distributions of the delivery demand and continually updated during the service time to leverage the crowd participation as needed. To address the resulting complex network management problem, we derive analytical solutions for compensation optimization and use these results along with effective pruning strategies to build a lookup table to simultaneously determine package routes and compensation offers in real time. Computational studies and extensive simulations conducted with real-world data show that our proposed approach can provide significant cost savings and considerably reduce operational costs and other transport- related negative externalities when compared to classical delivery modes, crowdshipping with static compensations, and crowdshipping without transshipment.
