Table 5 Parameter settings.

From: NF-MORL: a neuro-fuzzy multi-objective reinforcement learning framework for task scheduling in fog computing environments

Parameter

Value/configuration

Learning rate (Actor)

3 × 10⁻⁴

Learning rate (Critic)

1 × 10⁻³

Learning rate (fuzzy layer)

1 × 10⁻³

Discount factor (γ)

0.97

Entropy coefficient (β)

0.002

Batch size

64

Max steps per episode

1000

Number of episodes

150

Optimizer

Adam

Hidden layers (actor/critic)

2 × 256 neurons

Activation function

ReLU (hidden), SoftMax (output)

Reward weights (α₁, α₂, α₃, α₄)

(0.35, 0.25, 0.25, 0.15)

Fuzzy rules

81 (adaptive Takagi–Sugeno)

Back to article page