Agent-as-Tool: A Study on the Hierarchical Decision Making with Reinforcement Learning

Large Language Models (LLMs) have emerged as one of the most significant technological advancements in artificial intelligence in recent years. Their ability to understand, generate, and reason with natural language has transformed how we interact with AI systems. With the development of LLM-based agents and reinforcement-learning-based reasoning models, the study of applying reinforcement learning in agent frameworks has become a new research focus. However, all previous studies face the challenge of deciding the tool calling process and the reasoning process simultaneously, and the chain of reasoning was solely relied on the unprocessed raw result with redundant information and symbols unrelated to the task from the tool, which impose a heavy burden on the model's capability to reason. Therefore, in our research, we proposed a hierarchical framework Agent-as-tool that detach the tool calling process and the reasoning process, which enables the model to focus on the verbally reasoning process while the tool calling process is handled by another agent. Our work had achieved comparable results with only a slight reinforcement fine-tuning on 180 samples, and had achieved exceptionally well performance in Bamboogle with 63.2% of exact match and 75.2% in cover exact match, exceeding Search-R1 by 4.8% in exact match and 3.2% in cover exact match.

Strict Subgoal Execution: Reliable Long-Horizon Planning in Hierarchical Reinforcement Learning

Long-horizon goal-conditioned tasks pose fundamental challenges for reinforcement learning (RL), particularly when goals are distant and rewards are sparse. While hierarchical and graph-based methods offer partial solutions, they often suffer from subgoal infeasibility and inefficient planning. We introduce Strict Subgoal Execution (SSE), a graph-based hierarchical RL framework that enforces single-step subgoal reachability by structurally constraining high-level decision-making. To enhance exploration, SSE employs a decoupled exploration policy that systematically traverses underexplored regions of the goal space. Furthermore, a failure-aware path refinement, which refines graph-based planning by dynamically adjusting edge costs according to observed low-level success rates, thereby improving subgoal reliability. Experimental results across diverse long-horizon benchmarks demonstrate that SSE consistently outperforms existing goal-conditioned RL and hierarchical RL approaches in both efficiency and success rate.

Tailored Conversations beyond LLMs: A RL-Based Dialogue Manager

In this work, we propose a novel framework that integrates large language models (LLMs) with an RL-based dialogue manager for open-ended dialogue with a specific goal. By leveraging hierarchical reinforcement learning to model the structured phases of dialogue and employ meta-learning to enhance adaptability across diverse user profiles, our approach enhances adaptability and efficiency, enabling the system to learn from limited data, transition fluidly between dialogue phases, and personalize responses to heterogeneous patient needs. We apply our framework to Motivational Interviews, aiming to foster behavior change, and demonstrate that the proposed dialogue manager outperforms a state-of-the-art LLM baseline in terms of reward, showing a potential benefit of conditioning LLMs to create open-ended dialogue systems with specific goals.

Flow-Based Single-Step Completion for Efficient and Expressive Policy Learning

Generative models such as diffusion and flow-matching offer expressive policies for offline reinforcement learning (RL) by capturing rich, multimodal action distributions, but their iterative sampling introduces high inference costs and training instability due to gradient propagation across sampling steps. We propose the \textit{Single-Step Completion Policy} (SSCP), a generative policy trained with an augmented flow-matching objective to predict direct completion vectors from intermediate flow samples, enabling accurate, one-shot action generation. In an off-policy actor-critic framework, SSCP combines the expressiveness of generative models with the training and inference efficiency of unimodal policies, without requiring long backpropagation chains. Our method scales effectively to offline, offline-to-online, and online RL settings, offering substantial gains in speed and adaptability over diffusion-based baselines. We further extend SSCP to goal-conditioned RL, enabling flat policies to exploit subgoal structures without explicit hierarchical inference. SSCP achieves strong results across standard offline RL and behavior cloning benchmarks, positioning it as a versatile, expressive, and efficient framework for deep RL and sequential decision-making.

Efficient Navigation Among Movable Obstacles using a Mobile Manipulator via Hierarchical Policy Learning

We propose a hierarchical reinforcement learning (HRL) framework for efficient Navigation Among Movable Obstacles (NAMO) using a mobile manipulator. Our approach combines interaction-based obstacle property estimation with structured pushing strategies, facilitating the dynamic manipulation of unforeseen obstacles while adhering to a pre-planned global path. The high-level policy generates pushing commands that consider environmental constraints and path-tracking objectives, while the low-level policy precisely and stably executes these commands through coordinated whole-body movements. Comprehensive simulation-based experiments demonstrate improvements in performing NAMO tasks, including higher success rates, shortened traversed path length, and reduced goal-reaching times, compared to baselines. Additionally, ablation studies assess the efficacy of each component, while a qualitative analysis further validates the accuracy and reliability of the real-time obstacle property estimation.

HiLight: A Hierarchical Reinforcement Learning Framework with Global Adversarial Guidance for Large-Scale Traffic Signal Control

Efficient traffic signal control (TSC) is essential for mitigating urban congestion, yet existing reinforcement learning (RL) methods face challenges in scaling to large networks while maintaining global coordination. Centralized RL suffers from scalability issues, while decentralized approaches often lack unified objectives, resulting in limited network-level efficiency. In this paper, we propose HiLight, a hierarchical reinforcement learning framework with global adversarial guidance for large-scale TSC. HiLight consists of a high-level Meta-Policy, which partitions the traffic network into subregions and generates sub-goals using a Transformer-LSTM architecture, and a low-level Sub-Policy, which controls individual intersections with global awareness. To improve the alignment between global planning and local execution, we introduce an adversarial training mechanism, where the Meta-Policy generates challenging yet informative sub-goals, and the Sub-Policy learns to surpass these targets, leading to more effective coordination. We evaluate HiLight across both synthetic and real-world benchmarks, and additionally construct a large-scale Manhattan network with diverse traffic conditions, including peak transitions, adverse weather, and holiday surges. Experimental results show that HiLight exhibits significant advantages in large-scale scenarios and remains competitive across standard benchmarks of varying sizes.

Learning Swing-up Maneuvers for a Suspended Aerial Manipulation Platform in a Hierarchical Control Framework

In this work, we present a novel approach to augment a model-based control method with a reinforcement learning (RL) agent and demonstrate a swing-up maneuver with a suspended aerial manipulation platform. These platforms are targeted towards a wide range of applications on construction sites involving cranes, with swing-up maneuvers allowing it to perch at a given location, inaccessible with purely the thrust force of the platform. Our proposed approach is based on a hierarchical control framework, which allows different tasks to be executed according to their assigned priorities. An RL agent is then subsequently utilized to adjust the reference set-point of the lower-priority tasks to perform the swing-up maneuver, which is confined in the nullspace of the higher-priority tasks, such as maintaining a specific orientation and position of the end-effector. Our approach is validated using extensive numerical simulation studies.

Light Aircraft Game : Basic Implementation and training results analysis

This paper investigates multi-agent reinforcement learning (MARL) in a partially observable, cooperative-competitive combat environment known as LAG. We describe the environment's setup, including agent actions, hierarchical controls, and reward design across different combat modes such as No Weapon and ShootMissile. Two representative algorithms are evaluated: HAPPO, an on-policy hierarchical variant of PPO, and HASAC, an off-policy method based on soft actor-critic. We analyze their training stability, reward progression, and inter-agent coordination capabilities. Experimental results show that HASAC performs well in simpler coordination tasks without weapons, while HAPPO demonstrates stronger adaptability in more dynamic and expressive scenarios involving missile combat. These findings provide insights into the trade-offs between on-policy and off-policy methods in multi-agent settings.

Graph-Assisted Stitching for Offline Hierarchical Reinforcement Learning

Existing offline hierarchical reinforcement learning methods rely on high-level policy learning to generate subgoal sequences. However, their efficiency degrades as task horizons increase, and they lack effective strategies for stitching useful state transitions across different trajectories. We propose Graph-Assisted Stitching (GAS), a novel framework that formulates subgoal selection as a graph search problem rather than learning an explicit high-level policy. By embedding states into a Temporal Distance Representation (TDR) space, GAS clusters semantically similar states from different trajectories into unified graph nodes, enabling efficient transition stitching. A shortest-path algorithm is then applied to select subgoal sequences within the graph, while a low-level policy learns to reach the subgoals. To improve graph quality, we introduce the Temporal Efficiency (TE) metric, which filters out noisy or inefficient transition states, significantly enhancing task performance. GAS outperforms prior offline HRL methods across locomotion, navigation, and manipulation tasks. Notably, in the most stitching-critical task, it achieves a score of 88.3, dramatically surpassing the previous state-of-the-art score of 1.0. Our source code is available at: https://github.com/qortmdgh4141/GAS.

SkillBlender: Towards Versatile Humanoid Whole-Body Loco-Manipulation via Skill Blending

Humanoid robots hold significant potential in accomplishing daily tasks across diverse environments thanks to their flexibility and human-like morphology. Recent works have made significant progress in humanoid whole-body control and loco-manipulation leveraging optimal control or reinforcement learning. However, these methods require tedious task-specific tuning for each task to achieve satisfactory behaviors, limiting their versatility and scalability to diverse tasks in daily scenarios. To that end, we introduce SkillBlender, a novel hierarchical reinforcement learning framework for versatile humanoid loco-manipulation. SkillBlender first pretrains goal-conditioned task-agnostic primitive skills, and then dynamically blends these skills to accomplish complex loco-manipulation tasks with minimal task-specific reward engineering. We also introduce SkillBench, a parallel, cross-embodiment, and diverse simulated benchmark containing three embodiments, four primitive skills, and eight challenging loco-manipulation tasks, accompanied by a set of scientific evaluation metrics balancing accuracy and feasibility. Extensive simulated experiments show that our method significantly outperforms all baselines, while naturally regularizing behaviors to avoid reward hacking, resulting in more accurate and feasible movements for diverse loco-manipulation tasks in our daily scenarios. Our code and benchmark will be open-sourced to the community to facilitate future research. Project page: https://usc-gvl.github.io/SkillBlender-web/.