HiMAP-Travel: Hierarchical Multi-Agent Planning for Long-Horizon Constrained Travel

TL;DR

HiMAP-Travel introduces a hierarchical multi-agent planning framework that effectively manages long-horizon constrained travel tasks by coordinating parallel agents with global constraints, significantly improving success rates and efficiency.

Contribution

This work presents a novel hierarchical multi-agent framework with mechanisms for constraint enforcement and re-planning, advancing long-horizon planning capabilities in constrained environments.

Findings

Achieves over 52% validation and test Final Pass Rate on TravelPlanner.

Outperforms baseline methods like DeepTravel, ATLAS, and MTP significantly.

Reduces planning latency by 2.5 times through parallelization.

Abstract

Sequential LLM agents fail on long-horizon planning with hard constraints like budgets and diversity requirements. As planning progresses and context grows, these agents drift from global constraints. We propose HiMAP-Travel, a hierarchical multi-agent framework that splits planning into strategic coordination and parallel day-level execution. A Coordinator allocates resources across days, while Day Executors plan independently in parallel. Three key mechanisms enable this: a transactional monitor enforcing budget and uniqueness constraints across parallel agents, a bargaining protocol allowing agents to reject infeasible sub-goals and trigger re-planning, and a single policy trained with GRPO that powers all agents through role conditioning. On TravelPlanner, HiMAP-Travel with Qwen3-8B achieves 52.78% validation and 52.65% test Final Pass Rate (FPR). In a controlled comparison with identical model, training, and tools, it outperforms the sequential DeepTravel baseline by +8.67~pp. It also surpasses ATLAS by +17.65~pp and MTP by +10.0~pp. On FlexTravelBench multi-turn scenarios, it achieves 44.34% (2-turn) and 37.42% (3-turn) FPR while reducing latency 2.5x through parallelization.