An Intelligent Hybrid Framework Integrating PERT/CPM, COCOMO, SLIM, Machine Learning, and Multi-Agent Systems for Software Project Schedule Planning and Risk Prediction

The main contributions of this paper are:
1. A novel three-layer hybrid framework integrating traditional algorithmic models, ML, and MAS.
2. Empirical evaluation using four real-world datasets: NASA93, Desharnais, Maxwell, and Zenodo.
3. A four-agent MAS architecture with rule-based reasoning for dynamic buffer optimization.
4. Explainability analysis using SHAP to identify influential risk factors.