Research Article
Model-Driven Approach to Fading-Aware Wireless Network Planning Leveraging Multiobjective Optimization and Deep Learning
Table 4
Experiments for model-driven generation of predictive modules.
| Prediction module | Number of features | Number of layers | Manual writing (min) | Model creation (min) | Code generation (s) | Speed-up (times) |
| Service demand | 5 | 5 | 41 | 3.5 | 0.0138 | 11.71 | Energy consumption | 2 | 4 | 32 | 3 | 0.0101 | 10.27 |
|
|