use MarkovProcess in consumption saving models, fixes #639

Documentation/CHANGELOG.md

@@ -19,7 +19,7 @@ Release Data: TBD
 * calcExpectations method for taking the expectation of a distribution over a function [#884](https://github.com/econ-ark/HARK/pull/884/] (#897)[https://github.com/econ-ark/HARK/pull/897/)
 * Centralizes the definition of value, marginal value, and marginal marginal value functions that use inverse-space
 interpolation for problems with CRRA utility. See [#888](https://github.com/econ-ark/HARK/pull/888).
-* MarkovProcess class [#902](https://github.com/econ-ark/HARK/pull/902)
+* MarkovProcess class used in ConsMarkovModel, ConsRepAgentModel, ConsAggShockModel [#902](https://github.com/econ-ark/HARK/pull/902) [#929](https://github.com/econ-ark/HARK/pull/929)
 * Adds a SSA life tables and methods to extract survival probabilities from them [#986](https://github.com/econ-ark/HARK/pull/906).
 * Fix the return fields of `dcegm/calcCrossPoints`[#909](https://github.com/econ-ark/HARK/pull/909).
 * Corrects location of constructor documentation to class string for Sphinx rendering [#908](https://github.com/econ-ark/HARK/pull/908)

HARK/ConsumptionSaving/ConsAggShockModel.py

@@ -31,6 +31,7 @@
     makeGridExpMult,
 )
 from HARK.distribution import (
+    MarkovProcess,
     MeanOneLogNormal,
     Uniform,
     combineIndepDstns,
@@ -2547,7 +2548,6 @@ def makeMrkvHist(self):
 
         # Initialize the Markov history and set up transitions
         MrkvNow_hist = np.zeros(self.act_T_orig, dtype=int)
-        cutoffs = np.cumsum(self.MrkvArray, axis=1)
         loops = 0
         go = True
         MrkvNow = self.sow_init["MrkvNow"]
@@ -2557,9 +2557,10 @@ def makeMrkvHist(self):
         # Add histories until each state has been visited at least state_T_min times
         while go:
             draws = Uniform(seed=loops).draw(N=self.act_T_orig)
-            for s in range(draws.size):  # Add act_T_orig more periods
+            markov_process = MarkovProcess(self.MrkvArray,seed=loops)
+            for s in range(self.act_T_orig):  # Add act_T_orig more periods
                 MrkvNow_hist[t] = MrkvNow
-                MrkvNow = np.searchsorted(cutoffs[MrkvNow, :], draws[s])
+                MrkvNow = markov_process.draw(MrkvNow)
                 t += 1
 
             # Calculate the empirical distribution
@@ -2905,15 +2906,16 @@ def makeMrkvHist(self):
         """
         # Initialize the Markov history and set up transitions
         MrkvNow_hist = np.zeros(self.act_T, dtype=int)
-        cutoffs = np.cumsum(self.MrkvArray, axis=1)
         MrkvNow = self.MrkvNow_init
         t = 0
 
         # Add histories until each state has been visited at least state_T_min times
         draws = Uniform(seed=0).draw(N=self.act_T)
-        for s in range(draws.size):  # Add act_T_orig more periods
+
+        markov_process = MarkovProcess(self.MrkvArray, seed= 0)
+        for s in range(self.act_T):  # Add act_T_orig more periods
             MrkvNow_hist[t] = MrkvNow
-            MrkvNow = np.searchsorted(cutoffs[MrkvNow, :], draws[s])
+            MrkvNow = markov_process.draw(MrkvNow)
             t += 1
 
         # Store the result as attribute of self

HARK/ConsumptionSaving/ConsMarkovModel.py

@@ -19,6 +19,7 @@
 
 from HARK.distribution import (
     DiscreteDistribution,
+    MarkovProcess,
     Uniform,
     calcExpectation
 )
@@ -1057,15 +1058,6 @@ def getMarkovStates(self):
         -------
         None
         """
-        # Draw random numbers that will be used to determine the next Markov state
-        if self.global_markov:
-            base_draws = np.ones(self.AgentCount) * Uniform(
-                seed=self.RNG.randint(0, 2 ** 31 - 1)
-            ).draw(1)
-        else:
-            base_draws = Uniform(seed=self.RNG.randint(0, 2 ** 31 - 1)).draw(
-                self.AgentCount
-            )
         dont_change = (
             self.t_age == 0
         )  # Don't change Markov state for those who were just born (unless global_markov)
@@ -1076,20 +1068,15 @@ def getMarkovStates(self):
         J = self.MrkvArray[0].shape[0]
         MrkvPrev = self.shocks["MrkvNow"]
         MrkvNow = np.zeros(self.AgentCount, dtype=int)
-        MrkvBoolArray = np.zeros((J, self.AgentCount))
-
-        for j in range(J):
-            MrkvBoolArray[j, :] = MrkvPrev == j
 
         # Draw new Markov states for each agent
         for t in range(self.T_cycle):
-            Cutoffs = np.cumsum(self.MrkvArray[t], axis=1)
+            markov_process = MarkovProcess(
+                self.MrkvArray[t],
+                seed=self.RNG.randint(0, 2 ** 31 - 1)
+                )
             right_age = self.t_cycle == t
-            for j in range(J):
-                these = np.logical_and(right_age, MrkvBoolArray[j, :])
-                MrkvNow[these] = np.searchsorted(
-                    Cutoffs[j, :], base_draws[these]
-                ).astype(int)
+            MrkvNow[right_age] = markov_process.draw(MrkvPrev[right_age])
         if not self.global_markov:
             MrkvNow[dont_change] = MrkvPrev[dont_change]
 

HARK/ConsumptionSaving/ConsRepAgentModel.py

@@ -10,7 +10,7 @@
 from builtins import range
 import numpy as np
 from HARK.interpolation import LinearInterp, MargValueFuncCRRA
-from HARK.distribution import Uniform
+from HARK.distribution import (MarkovProcess, Uniform)
 from HARK.ConsumptionSaving.ConsIndShockModel import (
     IndShockConsumerType,
     ConsumerSolution,
@@ -349,12 +349,13 @@ def getShocks(self):
         -------
         None
         """
-        cutoffs = np.cumsum(self.MrkvArray[self.MrkvNow, :])
-        MrkvDraw = Uniform(seed=self.RNG.randint(0, 2 ** 31 - 1)).draw(N=1)
-        self.MrkvNow = np.searchsorted(cutoffs, MrkvDraw)
+        self.MrkvNow = MarkovProcess(
+            self.MrkvArray,
+            seed=self.RNG.randint(0, 2 ** 31 - 1)
+            ).draw(self.MrkvNow)
 
         t = self.t_cycle[0]
-        i = self.MrkvNow[0]
+        i = self.MrkvNow
         IncShkDstnNow = self.IncShkDstn[t - 1][i]  # set current income distribution
         PermGroFacNow = self.PermGroFac[t - 1][i]  # and permanent growth factor
         # Get random draws of income shocks from the discrete distribution
@@ -379,7 +380,7 @@ def getControls(self):
         None
         """
         t = self.t_cycle[0]
-        i = self.MrkvNow[0]
+        i = self.MrkvNow
         self.controls["cNrmNow"] = self.solution[t].cFunc[i](self.mNrmNow)