Coupling Planning with Episodic Memory in LLM Agents for Software Issue Resolution
Jiahao Zhang, Yifan Zhang, Yu Huang
Why It Matters
What makes this one worth your time
This research addresses the critical challenge of improving LLM agents' performance in software issue resolution, a task that is increasingly relevant as software systems grow in complexity.
PMCoder enhances software issue resolution by coupling planning with episodic memory.
Summary
The paper presents PMCoder, an issue-resolution agent that integrates a hierarchical phase planner with episodic memory to improve the resolution of software issues by maintaining an evolving plan and leveraging memory for better decision-making.
Key contributions
- Introduction of PMCoder, which couples hierarchical planning with episodic memory.
- Demonstration of improved issue resolution rates compared to baseline agents.
- Ablation studies showing the effectiveness of the coupling mechanism over isolated components.
Notable insights
- The bidirectional coupling between planning and memory allows for dynamic adaptation during the issue resolution process.
- Utilizing execution evidence for verification rather than self-reported claims enhances reliability in the agent's performance.
Possible limitations
- Not stated in the abstract.
Abstract
arXiv:2608.06811v1 Announce Type: cross Abstract: Resolving a real software issue with a large language model (LLM) agent is a long repair episode, often tens to hundreds of steps spanning exploration, hypothesis, implementation, and verification. Success depends on both the base model's local reasoning and the agent's ability to maintain an evolving plan and remember observations across phases. Existing repository-level agents typically strengthen planning or memory in isolation, leaving long trajectories vulnerable to stale evidence, repeated failed edits, and verification inferred from the agent's own claims instead of execution evidence. We present PMCoder, an issue-resolution agent that couples a hierarchical phase planner with episodic memory. The coupling is bidirectional: the current plan phase conditions memory retrieval, while memory-derived trajectory statistics inform stuck detection and replanning. When available, issue-reproduction verdicts ground verification progress in execution evidence rather than self-reported completion. On SWE-bench Verified, PMCoder resolves an average of $25$ more cases ($+5.0$pp) than a harness-matched baseline, with gains persisting even where the reproduction gate never fires. Further Verified-500 evaluations show the same positive direction across Claude Haiku 4.5, DeepSeek-V4-Flash, and an OpenHands port, with at least $14$ additional resolved cases ($+2.8$pp). Separately, evaluation on TerminalWorld's official sample suggests that the plan-memory substrate transfers beyond issue reports. Ablation and trajectory analyses show where the gains come from: coupling planning and memory outperforms either component alone and reduces repeated failed actions, empty-patch exits, and context-window exhaustion.