Table 3.

Algorithm parameter configuration specifications with final Values.

Algorithm component	Key parameters	Search space range	Final selected value	Optimization objective
CNN Layout Optimizer	Learning Rate	0.001–0.01	0.003	Minimize Spatial Conflicts
CNN Layout Optimizer	Batch Size	16–128	64	Minimize Spatial Conflicts
Multi-Objective Solver	Population Size	50–200	120	Maximize Pareto Efficiency
Multi-Objective Solver	Generations	100–500	300	Maximize Pareto Efficiency
Reinforcement Learning	Discount Factor (γ)	0.9–0.99	0.95	Maximize Long-term Reward
Reinforcement Learning	Exploration Rate (ε)	0.1–0.3	0.15	Maximize Long-term Reward
User Behavior Predictor	Hidden Units	64–512	256	Minimize Prediction Error
User Behavior Predictor	Dropout Rate	0.2–0.5	0.3	Minimize Prediction Error
Function Allocator	Clustering K	3–15	8	Maximize Functional Coherence
Function Allocator	Threshold	0.6–0.9	0.75	Maximize Functional Coherence
Real-time Scheduler	Update Frequency	1–10 Hz	5 Hz	Minimize Response Latency
Real-time Scheduler	Buffer Size	100–1000	500	Minimize Response Latency

Note: Final values were determined through grid search combined with Bayesian optimization over validation datasets from pilot deployments. .