LN: Li & Xu (2018) — Multi-Agent System for Distributed Multi-Project Scheduling
Bibliographic Reference
Citation: Li, F., & Xu, Z. (2018). A multi-agent system for distributed multi-project scheduling with two-stage decomposition. PLOS ONE, 13(10). https://doi.org/10.1371/journal.pone.0205445
Important Note
Overview
The bibliography lists “Confessore, G., Liotta, G., & Rismondo, S.” as the authors and cites a different title. The verified authors from the DOI are Feifeng Li and Zhe Xu. The verified title includes “with two-stage decomposition.” The DOI is correct but was paired with fabricated author metadata in the bibliography.
Pass 1 — Bird’s Eye View (5 Cs)
| C | Assessment |
|---|---|
| Category | System proposal + empirical evaluation |
| Context | Multi-agent approach to the classical multi-project scheduling problem |
| Correctness | Mathematical formulation + simulation evaluation. Standard OR methodology. |
| Contributions | (1) Two-stage decomposition: local project agents + global coordinator; (2) Negotiation-based resource conflict resolution; (3) Convergence proof |
| Clarity | Technical but well-structured. |
Relevance: ⭐⭐⭐
The “local agent + coordinator” architecture is the canonical design for PUMA Stage 5 multi-project PMO: individual project agents + master orchestrator.
PUMA Connection
The two-stage decomposition (local project agent handles own backlog, coordinator resolves resource conflicts) maps directly to PUMA Stage 5: each project has a specialist triage/estimation agent, while the Smart PMO orchestrates cross-project resource allocation.