Make sure to install docs requirements

Dockerfile

@@ -32,7 +32,7 @@ WORKDIR $CODE_DIR
 
 COPY --chown=${NB_UID}:${NB_GID} poetry.lock pyproject.toml ./
 
-RUN poetry export --no-interaction --no-ansi --without=dev -f requirements.txt --output requirements.txt \
+RUN poetry export --no-interaction --no-ansi --with=docs --without=dev -f requirements.txt --output requirements.txt \
     && pip install --no-cache-dir --requirement requirements.txt
 
 COPY --chown=${NB_UID}:${NB_GID} src/ src/

notebooks/nb_20_dynamic_programming.ipynb

@@ -446,6 +446,32 @@
     "plot_all_paths_graph(G, show_solution=True)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Value Iteration\n",
+    "\n",
+    "Another way to compute the optimal cost-to-go for all states that is also applicable in stochastic problems\n",
+    "is the **Value Iteration** algorithm:\n",
+    "\n",
+    "$$\n",
+    "\\begin{array}{l}\n",
+    "  \\textbf{Input}:\\ \\text{MDP}\\ M = \\langle S, s_0, A, P_a(s' \\mid s), r(s,a,s')\\rangle\\\\\n",
+    "  \\textbf{Output}:\\ \\text{Value function}\\ V\\\\[2mm]\n",
+    "  \\text{Set}\\ V\\ \\text{to arbitrary value function; e.g., }\\ V(s) = 0\\ \\text{for all}\\ s\\\\[2mm]\n",
+    "  \\text{repeat}\\ \\\\\n",
+    "  \\quad\\quad \\Delta \\leftarrow 0 \\\\\n",
+    "  \\quad\\quad \\text{foreach}\\ s \\in S \\\\\n",
+    "  \\quad\\quad\\quad\\quad \\underbrace{V'(s) \\leftarrow \\max_{a \\in A(s)} \\sum_{s' \\in S}  P_a(s' \\mid s)\\ [r(s,a,s') + \n",
+    " \\gamma\\ V(s') ]}_{\\text{Bellman equation}} \\\\\n",
+    "  \\quad\\quad\\quad\\quad \\Delta \\leftarrow \\max(\\Delta, |V'(s) - V(s)|) \\\\\n",
+    "  \\quad\\quad V \\leftarrow V' \\\\\n",
+    "  \\text{until}\\ \\Delta \\leq \\theta \n",
+    "\\end{array}\n",
+    "$$"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},