TY  - RPRT
TI  - HarnessBandit: Joint Learnability-Transferability Scheduling for Multi-Harness Agentic Reinforcement Learning
AU  - Hongliang Wei
AU  - Xiaobing Tu
AU  - Yinggui Wang
AU  - Zhengxi Liu
AU  - Rongkun Xue
AU  - Jinkui Ren
AU  - Xiantao Zhang
AU  - Debin Zhao
AU  - Xiaopeng Fan
PY  - 2026
UR  - https://arxiv.org/abs/2609.13739
ID  - 2609.13739
ER  -