Abstract
With the rapid development of the courier industry, customers are placing higher demands on the cost and delivery time of courier services. This paper addresses the bi-objective express transportation network design problem (BO-ETNDP) to minimize operation costs and maximum arrival times. A multi-structure parallel design methodology (MS-PDM) is proposed, evaluating seven common topological structures. For each structure, a novel bi-objective nonlinear mixed-integer optimization model is developed, incorporating hub sorting efficiency as a decision variable. A preference-based multi-objective algorithm (PB-MOA) is used to solve these models. Validated with data from a leading express company, the results demonstrate that the methodology avoids the limitations of single-structure optimization and identifies suitable topologies for various scenarios.
The Challenge: Balancing Speed and Cost
The global courier industry faces a persistent conflict between minimizing operational costs and meeting customer expectations for rapid delivery. Traditionally, network designers select a specific architecture—such as a "Hub-and-Spoke" system—and optimize within its constraints. However, this approach assumes the chosen structure is inherently superior, regardless of fluctuating business volumes.
Furthermore, standard models often treat hub sorting speed as a fixed parameter. In reality, companies can invest in advanced sorting equipment to reduce delivery times, making sorting efficiency a critical decision variable. This research addresses the Bi-Objective Express Transportation Network Design Problem (BO-ETNDP) by simultaneously optimizing network topology and sorting efficiency.
Methodology: Multi-Structure Parallel Design
The researchers introduced the Multi-Structure Parallel Design Methodology (MS-PDM). Rather than relying on a single network type, the framework evaluates seven distinct topological structures to find the optimal configuration for a given scenario.
The seven structures include:
- Fully Connected (FC): Direct point-to-point shipping between all nodes.
- Hub-and-Spoke (SAHS, MAHS, RAHS): Shipments are consolidated through central hubs with different allocation rules.
- Hybrid Structures (DSAHS, DMAHS, DRAHS): A combination of hub-and-spoke systems and direct lines for high-volume routes.
To solve these complex models, the team developed a Preference-Based Multi-Objective Algorithm (PB-MOA). This algorithm integrates ranking, branch-and-cut techniques, and Pareto dominance theory to identify solutions that balance cost and speed according to decision-maker preferences.
Case Study and Key Insights
The methodology was tested using real-world data from a major Chinese express company, involving 34 demand nodes and 6 candidate hubs. The analysis revealed several critical insights:
- Dynamic Topology: No single structure is universally optimal. As demand grows, the ideal network shifts. For instance, while hub-and-spoke systems are efficient for low volumes, the Fully Connected (FC) structure becomes cost-competitive and significantly faster when demand increases by over 20%.
- Sorting Efficiency as a Lever: Investing in faster hub sorting equipment can drastically reduce "hold time," improving overall network timeliness without the need for additional vehicles.
- The Hybrid Advantage: In the specific case study, the DRAHS (Direct-connection R-allocation hub-and-spoke) structure emerged as the most balanced solution. It utilized direct lines for high-traffic routes while maintaining hubs for lower-volume consolidation.
Results and Performance
The study highlighted the trade-offs inherent in different designs. The Fully Connected (FC) network provided the fastest delivery (44.21 hours) but at the highest cost (124.56 million Yuan). In contrast, the DMAHS structure was the most economical (109.07 million Yuan) but resulted in a maximum arrival time of approximately 89 hours. Ultimately, the DRAHS structure was selected as the optimal compromise, achieving an arrival time of 72.94 hours at a cost of 110.09 million Yuan.
Conclusion and Future Directions
While the MS-PDM framework offers a robust tool for network design, the authors note that future work should incorporate dynamic factors such as vehicle scheduling, traffic congestion, and seasonal demand spikes. Moving from a "snapshot" optimization to multi-period planning will further enhance the resilience and efficiency of global logistics networks.
Summary updated