Intelligent Solution System towards Parts Logistics Optimization

Yaoting Huang, Boyu Chen, Wenlian Lu, Zhong-Xiao Jin, Ren Zheng

Abstract

Due to the complication of the presented problem, intelligent algorithms show great power to solve the parts logistics optimization problem related to the vehicle routing problem (VRP). However, most of the existing research to VRP are incomprehensive and failed to solve a real-work parts logistics problem. In this work, towards SAIC logistics problem, we propose a systematic solution to this 2-Dimensional Loading Capacitated Multi-Depot Heterogeneous VRP with Time Windows by integrating diverse types of intelligent algorithms, including, a heuristic algorithm to initialize feasible logistics planning schemes by imitating manual planning, the core Tabu Search algorithm for global optimization, accelerated by a novel bundle technique, heuristically algorithms for routing, packing and queuing associated, and a heuristic post-optimization process to promote the optimal solution. Based on these algorithms, the SAIC Motor has successfully established an intelligent management system to give a systematic solution for the parts logistics planning, superior than manual planning in its performance, customizability and expandability.

1. Introduction: The 2L-VRPTW Challenge

In the automotive industry, logistics is a high-stakes jigsaw puzzle. Thousands of parts must be collected from diverse suppliers and delivered to assembly plants within strict time windows. This problem is a complex variant of the Vehicle Routing Problem (VRP), specifically identified as the 2-Dimensional Loading Capacitated Multi-Depot Heterogeneous VRP with Time Windows (2L-VRPTW).

The complexity arises from several real-world constraints:

  • 2D Loading Constraints: Parts are stored in bins that must be packed into a truck's floor space without overlapping and in a specific sequence to allow forklift access.
  • Heterogeneous Fleet: The logistics network utilizes trucks of varying sizes and models.
  • Multi-Depot & Time Windows: Trucks originate from various yards and must reach specific docks at suppliers or warehouses during narrow operational windows.
  • Queuing Dynamics: Simultaneous arrivals at a dock lead to queuing, introducing significant delays.

Traditional manual planning at SAIC Motor struggled to handle the combinatorial explosion of these constraints, necessitating a more robust, automated approach.

2. Methodology: The Three-Stage Intelligent Solution System

The proposed system, named SPRUCE, employs a three-stage approach to optimize the logistics network.

2.1 Initialization via Community Detection

To establish a feasible starting point, the system imitates manual logic through a community detection algorithm. This clusters suppliers geographically, and shipments are assigned to trucks in descending order based on distance. This "seeding" process ensures the optimization algorithm begins with a structured, feasible state rather than a chaotic one.

2.2 Core Optimization: Tabu Search with Shipment Bundling

The heart of the system is a Tabu Search (TS) algorithm. A primary challenge in large-scale VRP is the computational cost of moving individual shipments between routes. To address this, the authors introduced a shipment bundling technique. Instead of individual moves, the algorithm groups shipments from the same supplier to the same destination into "bundles." This reduction in complexity allows the algorithm to take larger steps in the solution space, effectively escaping local minima and accelerating convergence.

2.3 2L-VRPTW Solver and Post-optimization

An auxiliary solver manages 2D packing and route pruning. It utilizes a tree-search method for routing and a heuristic 2D packing algorithm (2PP) to ensure physical feasibility. Finally, a post-optimization stage performs "cleanup" tasks, such as replacing large trucks with smaller, more cost-effective alternatives or merging sub-routes to minimize unnecessary returns to the yard.

3. Results and Discussion

The system was evaluated using real-world industrial data from SAIC Motor, involving 311 shipments, 54 bin variants, 45 suppliers, and 8 plant warehouses.

3.1 Performance Gains

The impact of the intelligent system was substantial. The initial feasible scheme (mimicking manual planning) resulted in a total mileage of 1131.80. After applying the Tabu Search with bundling (TS-WB) and post-optimization, the total mileage dropped to 482.50—a 57% reduction.

Crucially, the bundling technique proved essential; without it (TS-NB), the algorithm stalled at a mileage of 1014.60, demonstrating that standard TS is insufficient for this level of constraint complexity.

3.2 Operational Efficiency

The SPRUCE system demonstrates high scalability, capable of processing 2,000 shipments in approximately 10 to 15 minutes. This performance makes it viable for daily operational planning in fast-paced manufacturing environments, with the potential to scale up to 10,000 daily shipments.

4. Conclusion

By integrating heuristic initialization, bundled Tabu Search, and specialized 2D packing solvers, the SPRUCE system provides a comprehensive solution to the 2L-VRPTW. It significantly outperforms manual planning in mileage reduction and operational efficiency. Future work will focus on further accelerating computational speed through parallelization and exploring Deep Q-learning for more advanced 2D packing heuristics.

Summary updated

Abstract

Due to the complication of the presented problem, intelligent algorithms show great power to solve the parts logistics optimization problem related to the vehicle routing problem (VRP). However, most of the existing research to VRP are incomprehensive and failed to solve a real-work parts logistics problem. In this work, towards SAIC logistics problem, we propose a systematic solution to this 2-Dimensional Loading Capacitated Multi-Depot Heterogeneous VRP with Time Windows by integrating diverse types of intelligent algorithms, including, a heuristic algorithm to initialize feasible logistics planning schemes by imitating manual planning, the core Tabu Search algorithm for global optimization, accelerated by a novel bundle technique, heuristically algorithms for routing, packing and queuing associated, and a heuristic post-optimization process to promote the optimal solution. Based on these algorithms, the SAIC Motor has successfully established an intelligent management system to give a systematic solution for the parts logistics planning, superior than manual planning in its performance, customizability and expandability.
BibTeX
@inproceedings{huang2019intelligent,
  author = {Yaoting Huang and Boyu Chen and Wenlian Lu and Zhong-Xiao Jin and Ren Zheng},
  title = {{Intelligent Solution System towards Parts Logistics Optimization}},
  booktitle = {WCGO},
  year = {2019},
  url = {https://openreview.net/forum?id=IVLfTpQZN2N},
}
Figure