Vehicle-as-Prompt: A Unified Deep Reinforcement Learning Framework for Heterogeneous Fleet Vehicle Routing Problem

2026-04-06

AI Summary

This work addresses the heightened sensitivity and computational complexity in solving the Heterogeneous Fleet Vehicle Routing Problem (HFVRP), which arises from diverse vehicle costs, varying capacities, and intricate real-world constraints. To tackle this challenge, the authors propose VaP-CSMV, a unified deep reinforcement learning framework that formulates HFVRP as a single-stage autoregressive decision process. The approach introduces a novel Vehicle-as-Prompt mechanism, integrating a cross-semantic encoder with a multi-view decoder to capture end-to-end the complex mappings between vehicle heterogeneity and customer attributes. Notably, the model achieves zero-shot generalization to large-scale instances and unseen problem variants without requiring instance-specific tuning. Empirical results demonstrate that VaP-CSMV significantly outperforms existing deep reinforcement learning methods on multiple HFVRP benchmarks, delivering solution quality comparable to classical heuristics while requiring only seconds of inference time.

Abstract

Unlike traditional homogeneous routing problems, the Heterogeneous Fleet Vehicle Routing Problem (HFVRP) involves heterogeneous fixed costs, variable travel costs, and capacity constraints, rendering solution quality highly sensitive to vehicle selection. Furthermore, real-world logistics applications often impose additional complex constraints, markedly increasing computational complexity. However, most existing Deep Reinforcement Learning (DRL)-based methods are restricted to homogeneous scenarios, leading to suboptimal performance when applied to HFVRP and its complex variants. To bridge this gap, we investigate HFVRP under complex constraints and develop a unified DRL framework capable of solving the problem across various variant settings. We introduce the Vehicle-as-Prompt (VaP) mechanism, which formulates the problem as a single-stage autoregressive decision process. Building on this, we propose VaP-CSMV, a framework featuring a cross-semantic encoder and a multi-view decoder that effectively addresses various problem variants and captures the complex mapping relationships between vehicle heterogeneity and customer node attributes. Extensive experimental results demonstrate that VaP-CSMV significantly outperforms existing state-of-the-art DRL-based neural solvers and achieves competitive solution quality compared to traditional heuristic solvers, while reducing inference time to mere seconds. Furthermore, the framework exhibits strong zero-shot generalization capabilities on large-scale and previously unseen problem variants, while ablation studies validate the vital contribution of each component.

Problem

• Heterogeneous Fleet Vehicle Routing Problem
• Deep Reinforcement Learning
• Vehicle Selection
• Complex Constraints
• Logistics Optimization

Innovation

• Vehicle-as-Prompt
• Heterogeneous Fleet Vehicle Routing Problem
• Deep Reinforcement Learning
• Cross-semantic Encoder
• Zero-shot Generalization