Factorization -Learning Initialization for Parameter Optimization in Cellular Networks

Table 1

Simulation parameters.

(a)

-learning parameters	Value
Learning rate	0.2
Discount factor	0.995
Number of states	46656
Number of actions	16
Initial exploration probability (-greedy)	1
Minimum exploration probability (-greedy)	0.05
Exploration decay rate (-greedy)	0.9995
(-learning)	0.5
Number of planning (Dyan -learning)	100

(b)

Factorization parameters	Value
Latent factor	8
Max iterations	500
Optimization algorithm	Stochastic gradient descent
Linear regularization
Regularization
Factorization threshold

(c)

Network parameters	Value
SINR target	0 dB
UE movement speed	2 km/h