Neural Deconstruction Search for Vehicle Routing Problems

2025-01-07

AI Summary

This work addresses the efficiency–quality trade-off bottleneck in solving Vehicle Routing Problems (VRPs). We propose the first neural combinatorial optimization framework grounded in a “deconstruct–reconstruct” paradigm—departing from conventional autoregressive solution construction. Our method employs reinforcement learning–driven neural policy iteration to deconstruct an initial solution, followed by dynamic reconstruction via coordinated greedy insertion and local search, thereby enabling tight coupling between neural models and lightweight operations research heuristics. To enhance policy generalization, we introduce a novel self-supervised representation of deconstruction states. Evaluated on three canonical VRP benchmarks, our approach consistently outperforms state-of-the-art operations research methods, achieving average solution quality improvements of 1.8%–3.2%. This marks a significant departure from—and advancement beyond—the limitations inherent in sequential construction paradigms.

Technology Category

• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Computer Vision: Learning & Optimization for CV

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Autoregressive construction approaches generate solutions to vehicle routing problems in a step-by-step fashion, leading to high-quality solutions that are nearing the performance achieved by handcrafted, operations research techniques. In this work, we challenge the conventional paradigm of sequential solution construction and introduce an iterative search framework where solutions are instead deconstructed by a neural policy. Throughout the search, the neural policy collaborates with a simple greedy insertion algorithm to rebuild the deconstructed solutions. Our approach surpasses the performance of state-of-the-art operations research methods across three challenging vehicle routing problems of various problem sizes.

Problem

• Efficient Routing
• Delivery Trucks
• Optimization

Innovation

• Neural Decomposition Search
• Recombination of Solutions
• Outperforms Heuristic Methods