Table 2 Hyperparameter settings for hybrid metaheuristic–MLP models.
Algorithm | Hyperparameter | Value/range | Description/reference |
|---|---|---|---|
BBO-MLP | Habitat size | 50–500 | Population size (common in BBO literature) |
Mutation probability | 0.01 | Typical value from BBO optimization studies | |
Migration probability | 0.7 | Standard exploration–exploitation balance | |
Number of generations | 1000 | Sufficient for convergence in preliminary tests | |
MVO-MLP | Universe size | 50–500 | Population size |
Wormhole existence probability (WEP) | 0.8 | Controls exploitation intensity | |
Traveling distance rate (TDR) | 1 | Standard recommended value | |
Number of iterations | 1000 | Ensures convergence | |
VS-MLP | Swarm size | 50–500 | Population size for the algorithm |
Social coefficient (c1) | 1.5 | Guides individual versus social learning | |
Cognitive coefficient (c2) | 1.5 | Standard for VS optimization | |
Max iterations | 1000 | Convergence criteria | |
BSA-MLP | Population size | 50–500 | Standard value in BSA applications |
Step size | 0.1 | Controls search granularity | |
Visual parameter | 0.2 | Determines neighborhood visibility | |
Iterations | 1000 | Ensures convergence of the search |
