---
title: "Improving Cross-Problem Vehicle Routing with Locally Augmented Preferences and Representation Disentanglement"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.24859v1"
permalink: "https://feed.astralor.com/items/itm_8bf58d1d05b7119874f8fd23"
published: "2026-08-25T17:44:23.000Z"
score: 75
channel: "ai"
tags: ["PLE","POLAR","Preference Optimization","Multi-task Learning","Representation Disentanglement","Vehicle Routing Problem"]
---

# Improving Cross-Problem Vehicle Routing with Locally Augmented Preferences and Representation Disentanglement

## AI commentary

> This is a noteworthy contribution to multi-task combinatorial optimization. POLAR and PLE target training signals and representation separation respectively, and their combination yields clear gains. It is worth reading for researchers in logistics algorithms and reinforcement learning.

## At a glance

Multi-task vehicle routing problem (VRP) solvers aim to handle multiple VRP variants within a single unified model, avoiding training a separate model for each variant. Existing methods face limitations in training and architecture: reinforcement learning suffers from reward-scale disparities and shrinking advantage signals, while preference optimization stagnates when sampled tours become near-identical; fully shared encoders entangle constraint-dependent representations. This paper proposes two model-agnostic contributions: POLAR, a training algorithm that applies a local search refinement pass to the best decoded tour before forming preference pairs, yielding more informative pairwise margins; and PLE, an encoder that routes each layer through one shared expert and task-specific experts via a gating mechanism, progressively separating common routing structure from constraint-specific encodings. Experiments on various VRP variants show POLAR and PLE together elevate the state-of-the-art among neural multi-task solvers, reducing the average gap to reference solutions by 21.3% relative to the strongest baseline on 16 in-distribution variants, and outperforming prior neural methods on 27 out of 32 unseen variants.

## Key points

- Proposes POLAR training algorithm that uses local search refinement to generate more informative preference pairs, mitigating signal decay in preference optimization
- Proposes PLE encoder architecture that routes through shared and task-specific experts via gating, disentangling common and constraint-specific representations
- Reduces average gap to reference solutions by 21.3% relative to strongest baseline on 16 in-distribution VRP variants, outperforming prior neural methods on 27 of 32 unseen variants
- Ablation studies confirm both contributions improve cross-problem generalization across multiple backbone architectures

## Full text

Multi-task vehicle routing problem (VRP) solvers seek to handle multiple VRP variants within a single unified model, avoiding the need to train a separate model for every variant. In spite of recent progress, current approaches remain limited on two fronts. On the training side, reinforcement learning suffers from reward-scale disparities and shrinking advantage signals as policies improve, whereas preference optimization stagnates once sampled tours become near-identical and thus fundamentally limited by the quality of the policy's own generated solutions, leaving both paradigms with weak supervision as training progresses. On the architecture side, existing fully shared encoders entangle constraint-dependent representations across heterogeneous variants, which limits generalization. We address these gaps with two model-agnostic contributions. First, we propose Preference Optimization with Locally Augmented Refinement (POLAR), a novel training algorithm that applies a local search refinement pass to the best decoded tour before forming preference pairs, yielding much more informative pairwise margins. Second, a Progressive Layered Extraction (PLE) encoder routes each encoder layer through one shared expert and a set of task-specific experts via a gating mechanism, progressively separating common routing structure from constraint-specific encodings. Through extensive experiments on various VRP variants, we show that POLAR and PLE together elevate the current state-of-the-art among neural multi-task solvers. We reduce the average gap to reference solutions by 21.3% relative to the strongest published baseline on 16 in-distribution variants, and outperform prior neural methods on 27 out of 32 unseen variants. Ablation studies confirm the efficacy of each contribution, showing that both improve cross-problem generalization across multiple backbone model architectures.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_8bf58d1d05b7119874f8fd23