td0 = td.LinearTDLambda(alpha=alpha, lam=lam, phi=phi, gamma=gamma) td0.name = r"TD({}) $\alpha$={}".format(lam, alpha) td0.color = "k" methods.append(td0) lam = 1. alpha = 0.004 mu = 0.0001 tdc = td.TDCLambda(alpha=alpha, mu=mu, lam=lam, phi=phi, gamma=gamma) tdc.name = r"TDC({}) $\alpha$={} $\mu$={}".format(lam, alpha, mu) tdc.color = "b" methods.append(tdc) alpha = .5 lam = 0.0 lstd = td.RecursiveLSPELambda(lam=lam, alpha=alpha, phi=phi, gamma=gamma) lstd.name = r"LSPE({}) $\alpha$={}".format(lam, alpha) lstd.color = "g" methods.append(lstd) lam = 0.0 eps = 100000 lstd = td.RecursiveLSTDLambda(lam=lam, eps=eps, phi=phi, gamma=gamma) lstd.name = r"LSTD({}) $\epsilon$={}".format(lam, eps) lstd.color = "g" lstd.ls = "-." methods.append(lstd) # alpha = .3 beta = 100. mins = 0
lam = .8 eps = 10000 lstd = td.RecursiveLSTDLambda(lam=lam, eps=eps, phi=phi) lstd.name = r"LSTD({})".format(lam) methods.append(lstd) lam = .0 eps = 100 lstd = td.RecursiveLSTDLambda(lam=lam, eps=eps, phi=phi) lstd.name = r"LSTD({})".format(lam) methods.append(lstd) lam = .8 alpha = 1. lspe = td.RecursiveLSPELambda(lam=lam, alpha=alpha, phi=phi) lspe.name = r"LSPE({}) $\alpha$={}".format(lam, alpha) methods.append(lspe) lam = .0 alpha = .01 beta = 1000 mins = 0 lstd = td.FPKF(lam=lam, alpha=alpha, beta=beta, mins=mins, phi=phi) lstd.name = r"FPKF({}) $\alpha={}$ $\beta={}$".format(lam, alpha, beta) lstd.ls = "--" methods.append(lstd) brm = td.RecursiveBRMDS(phi=phi) brm.name = "BRMDS" brm.color = "b"