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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hwang, Jaebak, Lee, Sanghyeon, Kim, Jeongmo, Han, Seungyul
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916029545840640
author Hwang, Jaebak
Lee, Sanghyeon
Kim, Jeongmo
Han, Seungyul
author_facet Hwang, Jaebak
Lee, Sanghyeon
Kim, Jeongmo
Han, Seungyul
contents 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, their reliance on conventional hindsight relabeling often fails to correct subgoal infeasibility, leading to inefficient high-level planning. To address this, we propose Strict Subgoal Execution (SSE), a graph-based hierarchical RL framework that integrates Frontier Experience Replay (FER) to separate unreachable from admissible subgoals and streamline high-level decision making. FER delineates the reachability frontier using failure and partial-success transitions, which identifies unreliable subgoals, increases subgoal reliability, and reduces unnecessary high-level decisions. Additionally, SSE employs a decoupled exploration policy to cover underexplored regions of the goal space and a path refinement that adjusts edge costs using observed low-level failures. Experimental results across diverse long-horizon benchmarks show that SSE consistently outperforms existing goal-conditioned and hierarchical RL methods in both efficiency and success rate. Our code is available at https://jaebak1996.github.io/SSE/
format Preprint
id arxiv_https___arxiv_org_abs_2506_21039
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Strict Subgoal Execution: Reliable Long-Horizon Planning in Hierarchical Reinforcement Learning
Hwang, Jaebak
Lee, Sanghyeon
Kim, Jeongmo
Han, Seungyul
Machine Learning
Artificial Intelligence
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, their reliance on conventional hindsight relabeling often fails to correct subgoal infeasibility, leading to inefficient high-level planning. To address this, we propose Strict Subgoal Execution (SSE), a graph-based hierarchical RL framework that integrates Frontier Experience Replay (FER) to separate unreachable from admissible subgoals and streamline high-level decision making. FER delineates the reachability frontier using failure and partial-success transitions, which identifies unreliable subgoals, increases subgoal reliability, and reduces unnecessary high-level decisions. Additionally, SSE employs a decoupled exploration policy to cover underexplored regions of the goal space and a path refinement that adjusts edge costs using observed low-level failures. Experimental results across diverse long-horizon benchmarks show that SSE consistently outperforms existing goal-conditioned and hierarchical RL methods in both efficiency and success rate. Our code is available at https://jaebak1996.github.io/SSE/
title Strict Subgoal Execution: Reliable Long-Horizon Planning in Hierarchical Reinforcement Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2506.21039