Research Article
Investigation of Effectiveness of Shuffled Frog-Leaping Optimizer in Training a Convolution Neural Network
Table 3
The initial parameters of algorithms.
| Algorithm | Parameter | Value |
| SFL | Maximum permitted change in a frog’s location | 10 | Number of memeplex | 20 | Number of frogs | 30 |
| ACO | Pheromone update constant (Q) | 15 | Global pheromone decay rate (pg) | 0.7 | Visibility sensitivity (β) | 7 | Population size | 70 | Number of ants | 15 | Maximum number of iterations | 35 | Local pheromone decay rate (pt) | 0.6 | Initial pheromone (τ) | 1e-06 | Pheromone sensitivity (α) | 1 | Pheromone constant (q) | 1 |
| BFSO | Probability of elimination | 0.1 | Spreading percentage %σ | 0.4 | Population size | 60 | Number of bacteria | 20 | Maximum number of iterations | 35 |
| WHO | Strategies | Decreasing the value of a | Whales attacking | Encircling | Max a | 5 | Probability of choosing spiral model | P ∈ [0, 1] | Probability of choosing shrinking encircling | p ∈ [0, 1] | Population size | 70 | Number of whales | 15 | Maximum number of iterations | 35 |
|
|