diaporamaMiscDM
 
◃  Ch. 6 App par renforcement  ▹
 

Calculs de Q(s,a) (1/2)

  • On pose β=0.2 et γ=0.4
  • Q(1,S) = 0 + β (1 + γ * 0 - 0 ) = 0.2
  • Q(1,E) = 0 + β (-1 + γ * 0 - 0) = -0,2
  • Q(4,N) = 0 + β (-1 + γ * 0.2 - 0) = 0.184
  • Q(2,O) = 0 + β (1 + γ * 0.2 -0) = 0,216
  • Q(4,S) = 0 + β (1 + γ * 0 - 0) = 0.2
  • Q(3,S) = -0.2  Q(0,N) = 0.2  Q(0,O) = -0.2
  • Q(5,E)= 0.2  Q(5,S)= -0.2
  • Q(7,N)= 0 + β (-1 + γ * 0.2 -0) = -0.184  Q(7,E)= 0.2