[Retracted] Application Method of Environmental Protection Building Elements Based on Artificial Intelligence Technology in the Field of Urban Planning and Design

Table 5

Experimental super parameters.

Super parameter name	Super parameter value
Batch size	128
Number of sampling rounds per step	20
Number of policy iterations per step	32
LSTM length	200
Experience pool capacity	100000
GAE estimation step	200
GAE	0.99
GAE	0.95
PPO cutoff factor	0.2
Value loss item weight	0.25
Entropy gain weight	0.01 → 0.001
Learning rate F	5e-5 → 1e-6
Adam	0.9
Adam	0.999