| MHMPA |
Cloud task scheduling |
Efficient computation; Pareto-optimal solutions |
Static model; lacks adaptability |
29.1% makespan reduction; energy and carbon savings (Abdel-Basset et al., 2021) |
| ML-MOO |
Adaptive dynamic scheduling |
Adjusts weights in real time; high adaptability |
High computational cost |
Responsive scheduling under dynamic demand (Pinki Kumar et al., 2025) |
| MIP |
CHP system optimization |
Accurate for constrained settings |
Sensitive to demand variation; complex computation |
7.5% system efficiency gain; 15% boost in self-consumption (Algieri, Morrone & Bova, 2020) |
| ML prediction model |
Renewable energy scheduling |
Nonlinear modeling; highly adaptive |
Requires high-quality data; weak on rare events |
87.5% simulation time cut; faster urban assessment (Li et al., 2022b) |
| Gradient-based optimization |
Logistics distribution |
Fast convergence; suitable for local optima |
Prone to local minima; poor global exploration |
12% transport cost cut; 9.5% carbon emission drop (Jiao & Zhang, 2025) |
| WSM |
Cost-load-renewable trade-off |
Computationally simple; handles structured trade-offs |
Highly weight-sensitive |
Total system cost up 15.1%; industrial +21.4%, residential +27.3% (Stoyanova & Monti, 2019; Zhao et al., 2023) |
