TL;DR
This paper tackles the challenge of optimizing urban logistics by proposing a novel bi-level programming model to select the best metro stations as distribution nodes for an underground logistics system (M-ULS). By considering both supplier and demander perspectives and leveraging the excess capacity of metro systems, the model aims to minimize total logistics costs while ensuring efficient and green transportation.
What Problem Does This Paper Address?
Urban areas worldwide are grappling with increasing traffic congestion, pollution, and the growing demands of e-commerce logistics. Traditional road-based transportation is becoming unsustainable, leading to a need for innovative solutions. Underground logistics systems (ULS), particularly those integrated with existing metro networks (M-ULS), offer a promising alternative. However, a critical challenge lies in determining the optimal locations for distribution nodes within the metro system to maximize efficiency and minimize costs. This paper addresses this by developing a model to select these nodes, considering the complex interplay between suppliers and demanders.
Methods
The core of this paper is a bi-level programming model. This approach is particularly suited for problems where decisions are made at different hierarchical levels, reflecting the distinct objectives of suppliers (providers) and demanders.
- Upper-Level Model: This level focuses on the provider's perspective. Its objective is to minimize the total cost of establishing and operating logistics distribution nodes. This includes construction costs, fixed costs, and transportation costs from suppliers to the nodes. The model considers investment budgets and node capacities.
- Lower-Level Model: This level represents the demander's perspective. Its objective is to minimize the cost of receiving goods from the selected distribution nodes. This model incorporates demand functions and reflects the demanders' selection behavior for logistics nodes.
A key aspect of the model is the application of symmetry principles to link the upper and lower levels. The authors also developed a heuristic algorithm implemented in MATLAB to solve this complex bi-level model, ensuring relatively fast convergence and optimality. The algorithm iteratively refines the selection of distribution nodes until a stable solution is reached.
Datasets
The study utilizes a case study based on Nanjing's Metro Line 2. Specific data points include:
- Metro Station Information: Coordinates and scale of stations along Metro Line 2.
- Alternative Logistics Distribution Nodes: Nine selected metro stations identified as potential distribution hubs.
- Transportation Costs: Unit transportation costs from alternative logistics distribution nodes to demanders.
- Distances: Distances from alternative logistics distribution nodes to demanders.
- Node Capacities and Costs: Maximum loading capacity, utility coefficients, investment fees, and storage/management fees for each potential distribution node.
- Supplier Information: Supply capacity, volume from supplier to nodes, distance from supplier to nodes, and unit transportation costs from supplier to nodes.
- Demanders: One provider (Maqun logistics park) and 19 requesters were considered in the case study.
No new benchmark datasets were introduced; the study relies on real-world data from Nanjing.
Key Insights and Ideas
- Metro Excess Capacity as a Logistics Resource: The paper highlights the significant potential of utilizing the underutilized capacity of metro systems during off-peak hours for freight transport, offering a sustainable and efficient alternative to road-based logistics.
- Bi-Level Optimization for Supplier-Demander Alignment: The bi-level programming model is crucial for capturing the conflicting yet interdependent objectives of logistics providers and demanders. This ensures that the selected node locations are not only cost-effective for providers but also meet the needs and preferences of demanders.
- Symmetry Principles in Logistics Network Design: The application of symmetry principles provides a structured way to link the decision-making processes at different levels of the logistics hierarchy, leading to a more integrated and robust solution.
- Node Reduction as an Outcome: The iterative solving process revealed that the optimal solution often involves a reduction in the number of active logistics distribution nodes, suggesting that consolidating operations at strategically chosen, highly efficient stations can lead to significant cost savings.
Key Results
The case study using Nanjing's Metro Line 2 demonstrated the feasibility and effectiveness of the proposed model. The iterative process converged to an optimal solution where six metro stations were selected as logistics distribution nodes. These optimal nodes are: Xianlinzhongxin station, Muxuyuan station, Xinjiekou station, Mochouhu station, Xinglongdajie station, and Yuantong station.
The total cost for the project was calculated to be 473,012,580 yuan. The goods delivered by each optimal node varied, with some nodes handling significantly larger volumes due to their operating cost advantages, even if their construction costs were higher. The model successfully identified locations that balanced supply and demand needs while minimizing overall logistics costs.
Limitations and Future Work
The authors acknowledge several limitations:
- Single Metro Line Focus: The study was limited to a single metro line (Nanjing Metro Line 2). Future work should investigate the complex metro system with multiple interconnected lines to further validate the model's applicability and reliability.
- Exclusion of Competition: The model did not consider the competitive relationship between new and existing logistics distribution nodes. This competition can influence service levels and should be incorporated in future research.
The authors suggest that further investigation into the competitive dynamics and the extension of the model to multi-line metro systems are important directions for future research.
Summary updated