def setAGrid(self):
if all([val for val in self.AReady.values()]):
self.A, self.prob = approx_markov(rho=self.rho,
sigma_u=self.sigma,
m=3,
n=self.Angp)
self.A = np.exp(self.A)
self.AMean = np.dot(solveStationary(self.prob), self.A)
self.KStarReady['AMean'] = True
self.setKStar()
def setsGrid(self):
if all([val for val in self.sReady.values()]):
self.s, self.prob = approx_markov(rho=self.rho,
sigma_u=self.sigma,
m=3,
n=self.sngp)
self.s = np.exp(self.s) * self.scale
self.sMean = np.dot(solveStationary(self.prob), self.s)
self.IReady['sMean'] = True
self.setIGrid()
self.etaReady['sMean'] = True
self.setEta()
def __init__(self,
gamma=1.5,
beta=0.97,
R=1.025,
rho=0.9,
sigma=0.1,
yngp=7,
WMax=100.0,
Wngp=10,
spline_order=3,
quad_order=5,
approach='value'):
self.gamma = gamma
self.beta = beta
self.R = R
self.approach = approach # approach should be one of 'value' or 'euler'
self.WMax = WMax
self.tol = 0.0001
self.cMin = 0.000001
self.rho = rho
self.sigma = sigma
self.yngp = yngp
# yProb[0,:] contains the transition probabilities for yGrid[0]
self.yGrid, self.yProb = approx_markov(rho=self.rho,
sigma_u=self.sigma,
m=3,
n=self.yngp)
self.yGrid = np.exp(self.yGrid)
self.yGridFine, self.yProbFine = approx_markov(rho=self.rho,
sigma_u=self.sigma,
m=3,
n=self.yngp * 4)
self.yGridFine = np.exp(self.yGridFine)
# E[Fy|y=y[0]] = yProb[0, :] * yGrid
self.Wngp = Wngp
self.WGrid = np.linspace(self.WMin, self.WMax, num=self.Wngp)
self.WGridFine = np.linspace(self.WMin, self.WMax, num=self.Wngp * 10)
self.spline_order = spline_order
self.quad_order = quad_order
def sigma_u(self, value):
self._sigma_u = value
self.isSet3['sigma_u'] = True
if all([val for val in self.isSet3.values()]):
#mc = tauchen(rho=self.rho, sigma_u=self.sigma_u, b=0., m=3, n=self.Angp)
#self.p = mc.P
#self.A = np.exp(mc.state_values)
self.A, self.p = approx_markov(rho=self.rho,
sigma_u=self.sigma_u,
m=3,
n=self.Angp)
self.A = np.exp(self.A)
self.AMean = np.dot(solveStationary(self.p), self.A)
def Angp(self, value):
self._Angp = value
self.isSet3['Angp'] = True
self.isSet4['Angp'] = True
if all([val for val in self.isSet3.values()]):
#mc = tauchen(rho=self.rho, sigma_u=self.sigma_u, b=0., m=3, n=self.Angp)
#self.p = mc.P
#self.A = np.exp(mc.state_values)
self.A, self.p = approx_markov(rho=self.rho,
sigma_u=self.sigma_u,
m=3,
n=self.Angp)
self.A = np.exp(self.A)
self.AMean = np.dot(solveStationary(self.p), self.A)
if all([val for val in self.isSet4.values()]):
self.currReturn = np.empty((self.Angp, self.kngp, self.kngp))
self.ixnBest = np.empty((self.Angp, self.kngp, self.kngp))
self.EV = np.zeros((self.Angp, self.kngp))
self.ixk1policy = np.empty((self.Angp, self.kngp), dtype=np.int32)
self.ixnpolicy = np.empty((self.Angp, self.kngp), dtype=np.int32)
def Sngp(self, value):
self._Sngp = value
self.isSet5['Sngp'] = True
self.isSet4['Sngp'] = True
if all([val for val in self.isSet5.values()]):
self.S, self.q = approx_markov(rho=self.phi,
sigma_u=self.sigma_v,
m=3,
n=self.Sngp)
self.S = np.exp(self.S)
self.SMean = np.dot(solveStationary(self.q), self.S)
if all([val for val in self.isSet4.values()]):
self.currReturn = np.empty(
(self.kngp, self.Angp, self.Sngp, self.kngp))
self.ixnBest = np.empty(
(self.kngp, self.Angp, self.Sngp, self.kngp))
self.EV = np.zeros((self.kngp, self.Angp, self.Sngp))
self.ixk1policy = np.empty((self.kngp, self.Angp, self.Sngp),
dtype=np.int32)
self.ixnpolicy = np.empty((self.kngp, self.Angp, self.Sngp),
dtype=np.int32)
def setwGrid(self):
if all([val for val in self.wReady.values()]):
self.w, self.prob = approx_markov(rho=0.0, sigma_u=self.sigma, m=3, n=self.wngp)
self.w = np.exp(self.w)
def setAGrid(self):
if all([val for val in self.AReady.values()]):
self.A, self.prob = approx_markov(rho=self.rho, sigma_u=self.sigma, m=3, n=self.Angp)
self.A = np.exp(self.A)
