Studi Tereksklusi (Excluded Studies)
Halaman ini mendokumentasikan daftar 10 studi yang dieksklusi pada tahap evaluasi kelayakan teks penuh (Full-Text Eligibility Screening / L2) Batch-1 beserta alasan eksklusi verbatim berdasarkan kriteria eksklusi (EC) yang ditetapkan dalam protokol SLR.
Ringkasan Alasan Eksklusi
Berdasarkan berkas SSOT 02_slr/extraction/DEF_excluded_batch1.csv, terdapat 10 studi yang dikeluarkan dari sintesis matriks utama:
- EC1 (Tidak ada teks penuh / Akses terbatas):
- B1-056: Tidak ada abstrak dan teks penuh tidak dapat diakses (akses berbayar Springer tanpa ketersediaan repositori institusional/VPN).
- B1-081: Teks penuh prosiding IEEE CNML 2026 terhalang WAF/akses tertutup. Bukti empiris primer dialihkan dan dievaluasi melalui naskah pengganti pra-cetak (preprint surrogate) B1-081P (S13).
- EC2 / IC2 (Studi sekunder / Tinjauan pustaka):
- B1-046: Studi tinjauan literatur sistematis (SLR); dialihkan sebagai benih penelusuran referensi balik/lanjut (snowballing seed).
- B1-083: Studi tinjauan literatur sistematis (SLR di IEEE Access); dialihkan ke penelusuran snowballing dan perumusan RQ4.
- B1-140: Studi tinjauan literatur multivokal (multivocal literature review terhadap 23 studi); dialihkan ke penelusuran snowballing serta diskursus RQ1/RQ4.
- EC2 / IC2 (Makalah posisi / Konseptual / Abstrak diperluas tanpa evaluasi empiris kuantitatif, QA < 5.0):
- B1-047: Makalah visi workshop (ASEW 2025) tanpa evaluasi tolok ukur empiris kuantitatif (QA 4.0 < 5.0).
- B1-051: Makalah kerja pendek (work-in-progress / extended abstract 5 halaman di FSE '25) tanpa validasi empiris (QA 3.5 < 5.0).
- B1-066: Makalah visi arsitektur pada pendamping konferensi (FSE Companion '25) tanpa implementasi konkret atau pengujian tolok ukur (QA 3.5 < 5.0).
- B1-074: Makalah konseptual/posisi LNCS tanpa evaluasi kuantitatif empiris (QA 4.0 < 5.0).
- B1-076: Makalah kerangka konseptual pada IEEE SoutheastCon 2025 tanpa pengujian empiris kuantitatif (QA 4.0 < 5.0).
Tabel Rincian 10 Studi Tereksklusi
| BID | Judul Studi | Tahun | Forum / Venue | QA | Keputusan | Alasan Eksklusi (Verbatim) | Sumber Data |
|---|---|---|---|---|---|---|---|
B1-046 |
LLM-Based Multi-Agent Systems for Software Engineering: Literature Review, Vision, and the Road Ahead | 2025 | ACM Transactions on Software Engineering and Methodology | 8.0 | EXCLUDE-L2 | EC2 / IC2 failure (Secondary study / Systematic Literature Review; retained as snowballing seed per screening protocol, not a primary empirical system proposal) | L2_R5.json |
B1-047 |
ALMAS: an Autonomous LLM-based Multi-Agent Software Engineering Framework | 2025 | IEEE/ACM International Conference on Automated Software Engineering Workshops (ASEW 2025) | 4.0 | EXCLUDE-L2 | EC2 / IC2 failure (Workshop vision / position paper without quantitative empirical benchmark evaluation; QA score 4.0 < 5.0) | L2_R5.json |
B1-051 |
Facilitating Trustworthy Human-Agent Collaboration in LLM-based Multi-Agent System oriented Software Engineering | 2025 | Companion Proceedings of the 33rd ACM Symposium on the Foundations of Software Engineering (FSE '25) | 3.5 | EXCLUDE-L2 | EC2 (Work-in-progress / extended abstract short paper, 5 pages, without empirical validation; QA score 3.5 < 5.0) | L2_R5.json |
B1-056 |
Multi-agent AI Framework for Developer Assistance: A New Paradigm in Software Engineering Automation | 2025 | Lecture notes in networks and systems (Springer chapter, closed) | 0.0 | EXCLUDE-L2 | EC1: no abstract/fulltext accessible (OpenAlex abs 0 words; Springer closed; user no institutional VPN for Springer); STOP per Two-Correction Rule | L2_R6.json |
B1-066 |
Knowledge-Based Multi-Agent Framework for Automated Software Architecture Design | 2025 | Companion Proceedings of the 33rd ACM International Conference on the Foundations of Software Engineering (FSE Companion '25) | 3.5 | EXCLUDE-L2 | EC2 / IC2 failure (Vision / position paper in FSE '25 Companion without empirical implementation or quantitative benchmark evaluation; QA score 3.5 < 5.0) | L2_R6.json |
B1-074 |
Developer Operations and Engineering Multi-agent Systems | — | Lecture Notes in Computer Science (LNCS, Springer) | 4.0 | EXCLUDE-L2 | EC2: conceptual/position paper, no empirical quantitative evaluation, QA 4.0<5.0 | L2_R7.json |
B1-076 |
A Multi-Agent LLM Environment for Software Design and Refactoring: A Conceptual Framework | 2025 | IEEE SoutheastCon 2025 (closed) | 4.0 | EXCLUDE-L2 | EC2: conceptual/position paper without empirical quantitative evaluation, QA 4.0<5.0 | L2_R7.json |
B1-081 |
Detecting and Repairing Role Drift in Multi-Agent Collaboration with Lightweight Protocols | 2026 | IEEE CNML 2026 (closed) | 7.0 | EXCLUDE-L2 | EC1: IEEE fulltext inaccessible (VPS WAF-blocked + user confirms no IEEE access 2026-09-10); abstract-only retained; primary evidence uses preprint surrogate B1-081P | L2_R8.json |
B1-083 |
Methods and Techniques of Agentic Software Engineering: A Systematic Literature Review | 2026 | IEEE Access 2026 (gold OA) | 7.0 | EXCLUDE-L2 | Secondary study: EXCLUDE from primary synthesis (like B1-046), moved to snowballing + RQ4 | L2_R8.json |
B1-140 |
Rethinking Software Engineering for Agentic AI Systems | 2026 | arXiv (cs.SE) | 7.5 | EXCLUDE-L2 | EC2: secondary study (multivocal literature review, 23 studies) - no novel empirical multi-agent system; moved to snowballing + RQ1/RQ4 discourse (cf. B1-046, B1-083) | L2_R10.json |