dorsal/arxiv
View SchemaStackPlanner: A Centralized Hierarchical Multi-Agent System with Task-Experience Memory Management
| Authors | Ruizhe Zhang, Xinke Jiang, Zhibang Yang, Zhixin Zhang, Jiaran Gao, Yuzhen Xiao, Hongbin Lai, Xu Chu, Junfeng Zhao, Yasha Wang |
|---|---|
| Categories | |
| ArXiv ID | 2601.05890vv1 |
| URL | https://arxiv.org/abs/2601.05890 |
| License | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
Abstract
Multi-agent systems based on large language models, particularly centralized architectures, have recently shown strong potential for complex and knowledge-intensive tasks. However, central agents often suffer from unstable long-horizon collaboration due to the lack of memory management, leading to context bloat, error accumulation, and poor cross-task generalization. To address both task-level memory inefficiency and the inability to reuse coordination experience, we propose StackPlanner, a hierarchical multi-agent framework with explicit memory control. StackPlanner addresses these challenges by decoupling high-level coordination from subtask execution with active task-level memory control, and by learning to retrieve and exploit reusable coordination experience via structured experience memory and reinforcement learning. Experiments on multiple deep-search and agent system benchmarks demonstrate the effectiveness of our approach in enabling reliable long-horizon multi-agent collaboration.
{
"annotation_id": "dd13677b-3731-4c2e-a756-bed4d8b55228",
"date_created": "2026-02-17T05:53:04.939000Z",
"date_modified": "2026-02-17T05:53:04.939000Z",
"file_hash": "5bfdc6e7b07c5f2150f91d6fd520b63466cf0da7da53829b2f7a61432332c800",
"private": false,
"record": {
"abstract": "Multi-agent systems based on large language models, particularly centralized architectures, have recently shown strong potential for complex and knowledge-intensive tasks. However, central agents often suffer from unstable long-horizon collaboration due to the lack of memory management, leading to context bloat, error accumulation, and poor cross-task generalization. To address both task-level memory inefficiency and the inability to reuse coordination experience, we propose StackPlanner, a hierarchical multi-agent framework with explicit memory control. StackPlanner addresses these challenges by decoupling high-level coordination from subtask execution with active task-level memory control, and by learning to retrieve and exploit reusable coordination experience via structured experience memory and reinforcement learning. Experiments on multiple deep-search and agent system benchmarks demonstrate the effectiveness of our approach in enabling reliable long-horizon multi-agent collaboration.",
"arxiv_id": "2601.05890",
"authors": [
"Ruizhe Zhang",
"Xinke Jiang",
"Zhibang Yang",
"Zhixin Zhang",
"Jiaran Gao",
"Yuzhen Xiao",
"Hongbin Lai",
"Xu Chu",
"Junfeng Zhao",
"Yasha Wang"
],
"categories": [
"cs.AI"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"title": "StackPlanner: A Centralized Hierarchical Multi-Agent System with Task-Experience Memory Management",
"url": "https://arxiv.org/abs/2601.05890",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "34c90d98-7658-48b2-afc9-6033d8224e27",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}