public MyStochasticReturnPredictor(MyRootDecisionSpace rds, int myPromotedVariable, MyModuleParams setup, String label, int level) : base(label, level, setup) { base.AddPromotedVariable(myPromotedVariable, rds); this.Rds = rds; Ds = new MyDecisionSpace(this, rds, setup); Mem = new MyQSAMemory(rds.VarManager.MAX_VARIABLES, 0); m_asm = new MyMotivationBasedDeleteUnselectedASM(setup); LearningAlgorithm = new MyDiscreteQLearning(setup, Mem); m_mlvh = new MyLocalVariableHistory(rds, m_setup, Ds); m_prevSelectedAction = 0; m_prev_st = Ds.GetCurrentState(); this.m_newVariables = new List <int>(); }
private int[] m_prev_st; // previous state (for variable adding and sharing knowledge) #endregion Fields #region Constructors public MyStochasticReturnPredictor(MyRootDecisionSpace rds, int myPromotedVariable, MyModuleParams setup, String label, int level) : base(label, level, setup) { base.AddPromotedVariable(myPromotedVariable, rds); this.Rds = rds; Ds = new MyDecisionSpace(this, rds, setup); Mem = new MyQSAMemory(rds.VarManager.MAX_VARIABLES, 0); m_asm = new MyMotivationBasedDeleteUnselectedASM(setup); LearningAlgorithm = new MyDiscreteQLearning(setup, Mem); m_mlvh = new MyLocalVariableHistory(rds, m_setup, Ds); m_prevSelectedAction = 0; m_prev_st = Ds.GetCurrentState(); this.m_newVariables = new List<int>(); }