Use JAX in successive_approx (#134)

Author: kp992

lectures/_static/lecture_specific/successive_approx.py

@@ -1,21 +1,21 @@
-def successive_approx(T,                     # Operator (callable)
-                      x_0,                   # Initial condition
-                      tolerance=1e-6,        # Error tolerance
-                      max_iter=10_000,       # Max iteration bound
-                      print_step=25,         # Print at multiples
-                      verbose=False):
-    x = x_0
-    error = tolerance + 1
-    k = 1
-    while error > tolerance and k <= max_iter:
-        x_new = T(x)
+def successive_approx_jax(x_0,                   # Initial condition
+                          constants,
+                          sizes,
+                          arrays,                 
+                          tolerance=1e-6,        # Error tolerance
+                          max_iter=10_000):      # Max iteration bound
+
+    def body_fun(k_x_err):
+        k, x, error = k_x_err
+        x_new = T(x, constants, sizes, arrays)
         error = jnp.max(jnp.abs(x_new - x))
-        if verbose and k % print_step == 0:
-            print(f"Completed iteration {k} with error {error}.")
-        x = x_new
-        k += 1
-    if error > tolerance:
-        print(f"Warning: Iteration hit upper bound {max_iter}.")
-    elif verbose:
-        print(f"Terminated successfully in {k} iterations.")
+        return k + 1, x_new, error
+
+    def cond_fun(k_x_err):
+        k, x, error = k_x_err
+        return jnp.logical_and(error > tolerance, k < max_iter)
+
+    k, x, error = jax.lax.while_loop(cond_fun, body_fun, (1, x_0, tolerance + 1))
     return x
+
+successive_approx_jax = jax.jit(successive_approx_jax, static_argnums=(2,))

lectures/_static/lecture_specific/vfi.py

@@ -2,8 +2,7 @@
 
 def value_iteration(model, tol=1e-5):
     constants, sizes, arrays = model
-    _T = lambda v: T(v, constants, sizes, arrays)
     vz = jnp.zeros(sizes)
 
-    v_star = successive_approx(_T, vz, tolerance=tol)
+    v_star = successive_approx_jax(vz, constants, sizes, arrays, tolerance=tol)
     return get_greedy(v_star, constants, sizes, arrays)

lectures/opt_invest.md

@@ -4,14 +4,13 @@ jupytext:
     extension: .md
     format_name: myst
     format_version: 0.13
-    jupytext_version: 1.14.5
+    jupytext_version: 1.16.1
 kernelspec:
   display_name: Python 3 (ipykernel)
   language: python
   name: python3
 ---
 
-
 # Optimal Investment
 
 ```{include} _admonition/gpu.md
@@ -76,14 +75,6 @@ We will use 64 bit floats with JAX in order to increase the precision.
 jax.config.update("jax_enable_x64", True)
 ```
 
-
-We need the following successive approximation function.
-
-```{code-cell} ipython3
-:load: _static/lecture_specific/successive_approx.py
-```
-
-
 Let's define a function to create an investment model using the given parameters.
 
 ```{code-cell} ipython3
@@ -113,7 +104,6 @@ def create_investment_model(
     return constants, sizes, arrays
 ```
 
-
 Let's re-write the vectorized version of the right-hand side of the
 Bellman equation (before maximization), which is a 3D array representing
 
@@ -183,7 +173,6 @@ def compute_r_σ(σ, constants, sizes, arrays):
 compute_r_σ = jax.jit(compute_r_σ, static_argnums=(2,))
 ```
 
-
 Define the Bellman operator.
 
 ```{code-cell} ipython3
@@ -194,7 +183,6 @@ def T(v, constants, sizes, arrays):
 T = jax.jit(T, static_argnums=(2,))
 ```
 
-
 The following function computes a v-greedy policy.
 
 ```{code-cell} ipython3
@@ -205,7 +193,6 @@ def get_greedy(v, constants, sizes, arrays):
 get_greedy = jax.jit(get_greedy, static_argnums=(2,))
 ```
 
-
 Define the $\sigma$-policy operator.
 
 ```{code-cell} ipython3
@@ -236,7 +223,6 @@ def T_σ(v, σ, constants, sizes, arrays):
 T_σ = jax.jit(T_σ, static_argnums=(3,))
 ```
 
-
 Next, we want to computes the lifetime value of following policy $\sigma$.
 
 This lifetime value is a function $v_\sigma$ that satisfies
@@ -285,8 +271,7 @@ def L_σ(v, σ, constants, sizes, arrays):
 L_σ = jax.jit(L_σ, static_argnums=(3,))
 ```
 
-Now we can define a function to compute $v_{\sigma}$ 
-
+Now we can define a function to compute $v_{\sigma}$
 
 ```{code-cell} ipython3
 def get_value(σ, constants, sizes, arrays):
@@ -306,6 +291,11 @@ def get_value(σ, constants, sizes, arrays):
 get_value = jax.jit(get_value, static_argnums=(2,))
 ```
 
+We use successive approximation for VFI.
+
+```{code-cell} ipython3
+:load: _static/lecture_specific/successive_approx.py
+```
 
 Finally, we introduce the solvers that implement VFI, HPI and OPI.
 
@@ -355,7 +345,6 @@ print(out)
 print(f"OPI completed in {elapsed} seconds.")
 ```
 
-
 Here's the plot of the Howard policy, as a function of $y$ at the highest and lowest values of $z$.
 
 ```{code-cell} ipython3
@@ -377,7 +366,6 @@ ax.legend(fontsize=12)
 plt.show()
 ```
 
-
 Let's plot the time taken by each of the solvers and compare them.
 
 ```{code-cell} ipython3
@@ -403,6 +391,7 @@ print(f"VFI completed in {vfi_time} seconds.")
 
 ```{code-cell} ipython3
 :tags: [hide-output]
+
 opi_times = []
 for m in m_vals:
     print(f"Running optimistic policy iteration with m={m}.")

lectures/opt_savings.md

@@ -4,7 +4,7 @@ jupytext:
     extension: .md
     format_name: myst
     format_version: 0.13
-    jupytext_version: 1.14.5
+    jupytext_version: 1.16.1
 kernelspec:
   display_name: Python 3 (ipykernel)
   language: python
@@ -65,20 +65,13 @@ where
 
 $$   u(c) = \frac{c^{1-\gamma}}{1-\gamma} $$
 
-+++
-
-We use successive approximation for VFI.
-
-```{code-cell} ipython3
-:load: _static/lecture_specific/successive_approx.py
-```
 
 ## Model primitives
 
 First we define a model that stores parameters and grids
 
 ```{code-cell} ipython3
-def create_consumption_model(R=1.01,                # Gross interest rate
+def create_consumption_model(R=1.01,                    # Gross interest rate
                              β=0.98,                    # Discount factor
                              γ=2,                       # CRRA parameter
                              w_min=0.01,                # Min wealth
@@ -140,8 +133,6 @@ which is defined as the vector
 
 $$ r_\sigma(w, y) := r(w, y, \sigma(w, y)) $$
 
-
-
 ```{code-cell} ipython3
 def compute_r_σ(σ, constants, sizes, arrays):
     """
@@ -187,9 +178,9 @@ def T_σ(v, σ, constants, sizes, arrays):
     Q = jnp.reshape(Q, (1, y_size, y_size))
 
     # Calculate the expected sum Σ_jp v[σ[i, j], jp] * Q[i, j, jp]
-    Ev = jnp.sum(V * Q, axis=2)
+    EV = jnp.sum(V * Q, axis=2)
 
-    return r_σ + β * Ev
+    return r_σ + β * EV
 ```
 
 and the Bellman operator $T$
@@ -260,7 +251,7 @@ def L_σ(v, σ, constants, sizes, arrays):
     return v - β * jnp.sum(V * Q, axis=2)
 ```
 
-Now we can define a function to compute $v_{\sigma}$ 
+Now we can define a function to compute $v_{\sigma}$
 
 ```{code-cell} ipython3
 def get_value(σ, constants, sizes, arrays):
@@ -291,6 +282,12 @@ T_σ = jax.jit(T_σ, static_argnums=(3,))
 L_σ = jax.jit(L_σ, static_argnums=(3,))
 ```
 
+We use successive approximation for VFI.
+
+```{code-cell} ipython3
+:load: _static/lecture_specific/successive_approx.py
+```
+
 ## Solvers
 
 Now we define the solvers, which implement VFI, HPI and OPI.
@@ -353,7 +350,7 @@ print("Starting VFI.")
 start_time = time.time()
 out = value_iteration(model)
 elapsed = time.time() - start_time
-print(f"VFI(jax not in succ) completed in {elapsed} seconds.")
+print(f"VFI completed in {elapsed} seconds.")
 ```
 
 ```{code-cell} ipython3