📖 MOC — Literature Review (SLR)
Overview
Systematic Literature Review for PUMA following PRISMA 2020 + PRISMA-DFLLM + PRISMA-trAIce protocols. Target: ≥ 40 papers · Period: 2022–2026 · SLR Workflow: WF-SLR-Pipeline
📊 PRISMA Status
| Stage | Count | Progress |
|---|---|---|
| Identified | ~[N] | ⏳ |
| Screened (title/abstract) | ~[N] | ⏳ |
| Eligible (full text) | ~[N] | ⏳ |
| Included | ≥40 | ⏳ |
Full log: PRISMA-Log
🗂️ Papers by Topic
Issue Triage & Priority Classification
| Paper | Year | Dataset | Metric | Notes |
|---|---|---|---|---|
| LN-Datasets-JiraSR-TAWOS | 2015 | Jira SR | — | Dataset paper (Ortu et al.) |
| LN-Chen-2025-AIOpsLab | 2025 | AIOps | F1 | AIOps benchmark lab |
| LN-MAGIS-2024-GitHubIssues | 2024 | GitHub | F1 | GitHub issue resolution via MAS |
| LN-Arora-2024-MASAI | 2024 | SWE-bench | Resolve% | Modular agent for SE |
| LN-Wang-2024-OpenHands | 2024 | SWE-bench | Resolve% | OpenHands coding agent |
| 20 - Literature/20.1 Papers/LN-Manzoor2025-AI-PM | 2025 | Multiple | Various | Survey (note not yet created) |
| [Add papers here as SLR progresses] |
Effort Estimation & Story Points
| Paper | Year | Dataset | Best MAE | Notes |
|---|---|---|---|---|
| LN-Datasets-JiraSR-TAWOS | 2022 | TAWOS | — | Dataset paper (Tawosi et al.) |
| LN-KeyPapers-CoGEE-Angermeir-Flyvbjerg | 2024 | TAWOS | ~1.9 SP | State of art (GPT-4) — CoGEE section |
| LN-Assalaarachchi-2026-AgenticSPM | 2026 | Various | — | Agentic SPM vision |
| LN-Cinkusz-2025-CognitiveAgentsAgilePM | 2025 | Various | — | Cognitive agents in Agile PM |
| LN-Li-2018-MultiProjectScheduling | 2018 | MRCPSP | — | Multi-project scheduling |
| 20 - Literature/20.1 Papers/LN-Yonathan2025-LocalLLMs | 2025 | TAWOS | ~3.2 SP | Local LLMs (note not yet created) |
| [Add papers here] |
LLM Benchmarks in Software Engineering
| Paper | Year | Tasks | Reproducible? | Notes |
|---|---|---|---|---|
| LN-Angermeir-2025-Reproducibility | 2025 | SE general | Meta-study | Reproducibility gap |
| LN-Jimenez-2023-SWEbench | 2023 | SWE-bench | Partial | SE coding benchmark |
| LN-Mialon-2023-GAIA | 2023 | GAIA | Partial | General AI assistant benchmark |
| LN-Hong-2023-MetaGPT | 2023 | HumanEval+ | Yes | Multi-role SE agents |
| 20 - Literature/20.1 Papers/LN-Berti2024-PM-LLM-Benchmark | 2024 | Process mining | Partial | PM+LLM (note not yet created) |
| [Add papers here] |
Prompting Strategies
| Paper | Year | Strategy | Task | Key finding |
|---|---|---|---|---|
| LN-Yao-2022-ReAct | 2022 | ReAct | Reasoning+Act | Combines reasoning with action |
| LN-Zelikman-2024-QuietSTaR | 2024 | Chain-of-thought | Reasoning | Implicit CoT rationales |
| LN-Calikli-2025-RequestFormats | 2025 | Multiple | Estimation | Non-monotonic effect |
| 20 - Literature/20.1 Papers/LN-Wei2022-CoT | 2022 | CoT | Reasoning | CoT helps ≥100B models (note not yet created) |
| 20 - Literature/20.1 Papers/LN-Brown2020-GPT3-FewShot | 2020 | Few-shot | Various | ICL discovery (note not yet created) |
| [Add papers here] |
Research Methodology
| Paper | Year | Method | Use in PUMA |
|---|---|---|---|
| LN-MITAILab-WP316-HowToDoResearch | 1988 | AI Lab method | Active reading + Q1/Q2/Q3 |
| LN-Spichkova-2025-CognitiveAgents | 2025 | Cognitive agents | Agent design patterns |
| 20 - Literature/20.1 Papers/LN-Hevner2004-DSR | 2004 | DSR | Paradigm (note not yet created) |
| 20 - Literature/20.1 Papers/LN-Peffers2007-DSRM | 2007 | DSRM | Process (note not yet created) |
| 20 - Literature/20.1 Papers/LN-Kitchenham2007-SLR | 2007 | SLR | Protocol (note not yet created) |
| 20 - Literature/20.1 Papers/LN-Page2021-PRISMA2020 | 2021 | PRISMA | Reporting (note not yet created) |
| 20 - Literature/20.1 Papers/LN-Wohlin2012-Experimentation | 2012 | Experiments | Design (note not yet created) |
🔍 Research Gap Summary
Based on SLR, three gaps define PUMA’s contribution:
| Gap | Evidence | PUMA Response |
|---|---|---|
| Reproducibility | 5/18 artefacts executable (Angermeir et al., 2025) | 100% local, seed=42, MIT |
| Prompting comparison | No systematic PM prompting study found | 4 strategies × 2 models |
| Carbon footprint | 0/N PM+LLM papers measure CO₂ | CodeCarbon per condition |
📌 Comparison Table (Extract)
TABLE authors, year, datasets_used, metrics, reproducible, puma_relevance
FROM "20 - Literature/20.1 Papers"
WHERE type = "literature-note" AND prisma_decision = "include"
SORT year DESC