Rethinking Light Decoder-based Solvers for Vehicle Routing Problems

2025-03-02

AI Summary

This paper addresses the weak generalization capability of lightweight decoder-based VRP solvers—particularly on large-scale instances and diverse VRP variants—identifying the root cause as static encoder overloading coupled with insufficient decoder capacity. To resolve this, we propose an enhanced lightweight decoder architecture: (i) incorporating identity mappings and additional feed-forward layers to boost decoder expressiveness without increasing encoder complexity; (ii) integrating a graph neural network (GNN) encoder with a reinforcement learning–based autoregressive decoding framework; and (iii) supporting the design with structured embedding analysis and systematic out-of-distribution (OOD) evaluation. Experiments demonstrate that our approach significantly narrows the performance gap with heavyweight decoder baselines across multiple VRP variants and large-scale scenarios, while markedly improving OOD generalization. The implementation is publicly available.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Sampling/Simulation-based Search

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: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Light decoder-based solvers have gained popularity for solving vehicle routing problems (VRPs) due to their efficiency and ease of integration with reinforcement learning algorithms. However, they often struggle with generalization to larger problem instances or different VRP variants. This paper revisits light decoder-based approaches, analyzing the implications of their reliance on static embeddings and the inherent challenges that arise. Specifically, we demonstrate that in the light decoder paradigm, the encoder is implicitly tasked with capturing information for all potential decision scenarios during solution construction within a single set of embeddings, resulting in high information density. Furthermore, our empirical analysis reveals that the overly simplistic decoder struggles to effectively utilize this dense information, particularly as task complexity increases, which limits generalization to out-of-distribution (OOD) settings. Building on these insights, we show that enhancing the decoder capacity, with a simple addition of identity mapping and a feed-forward layer, can considerably alleviate the generalization issue. Experimentally, our method significantly enhances the OOD generalization of light decoder-based approaches on large-scale instances and complex VRP variants, narrowing the gap with the heavy decoder paradigm. Our code is available at: https://github.com/ziweileonhuang/reld-nco.

Problem

• Improves generalization of light decoder-based VRP solvers.
• Addresses limitations in handling large-scale VRP instances.
• Enhances performance on complex VRP variants and OOD settings.

Innovation

• Enhanced decoder with identity mapping
• Added feed-forward layer for better generalization
• Improved OOD generalization in VRPs