{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Newsvendor Optimization\n", "Complete Enumeration, Inverting the CDF, Bisection Search, Derivative-free optimization with `scipy.optimize`" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "import matplotlib.pyplot as plt\n", "import pandas as pd\n", "from scipy import stats" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Generate Demand Data" ] }, { "cell_type": "code", "execution_count": 36, "metadata": {}, "outputs": [], "source": [ "# Use this alternate code for the holiday lights example in the lecture slides\n", "#demand = pd.DataFrame(np.exp(2*np.random.normal(size=10000)), columns=['demand'])\n", "\n", "# A discrete demand distribution\n", "mu = 50\n", "demand = pd.DataFrame(np.random.poisson(mu,size=1000), columns=['demand'])" ] }, { "cell_type": "code", "execution_count": 37, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([[]],\n", " dtype=object)" ] }, "execution_count": 37, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "demand.hist(bins=100)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Complete Enumeration" ] }, { "cell_type": "code", "execution_count": 38, "metadata": {}, "outputs": [], "source": [ "co = 10\n", "cu = 35\n", "maxq = 70\n", "\n", "# Expected cost of an inventory level q\n", "def cost(q):\n", " cost_overage = co*(q-demand).clip(lower=0).mean()[0]\n", " cost_underage = cu*(demand-q).clip(lower=0).mean()[0]\n", " return cost_overage + cost_underage" ] }, { "cell_type": "code", "execution_count": 40, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "55" ] }, "execution_count": 40, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Calculate where the cost is minimized via complete enumeration\n", "q_list = range(maxq)\n", "cost_list = [cost(q) for q in q_list]\n", "qstar = q_list[np.where(cost_list == np.min(cost_list))[0][0]]\n", "\n", "plt.plot(q_list,cost_list)\n", "plt.axvline(qstar,color='black')\n", " #plt.xlim(0,10)\n", " #plt.ylim(200,300)\n", "qstar" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Derivative" ] }, { "cell_type": "code", "execution_count": 41, "metadata": {}, "outputs": [], "source": [ "def cdf(q):\n", " return (demand <= q).mean()[0] # P(D <= Q)\n", "\n", "def derivative(q):\n", " # This is the derivative we calculated in the slides\n", " return -cu + (cu + co) * cdf(q) " ] }, { "cell_type": "code", "execution_count": 42, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 42, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# The optimal point is where the derivative is 0\n", "q_list = range(maxq)\n", "grad_list = [derivative(q) for q in q_list]\n", "plt.plot(q_list,grad_list)\n", "plt.axvline(qstar,color='black')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Approximate Q* by inverting the cdf" ] }, { "cell_type": "code", "execution_count": 43, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.7777777777777778" ] }, "execution_count": 43, "metadata": {}, "output_type": "execute_result" } ], "source": [ "a = cu / (cu + co) # Newsvendor quantile\n", "a" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "4.615496045093771" ] }, "execution_count": 21, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# First example, where the log of demand is normal with standard devation 2 and mean 0.\n", "# To handle a general non-zero mean of m and standard deviation s, use stats.lognorm(s,scale=np.exp(m))\n", "stats.lognorm(s=2).ppf(a)" ] }, { "cell_type": "code", "execution_count": 27, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "247.53531520516498" ] }, "execution_count": 27, "metadata": {}, "output_type": "execute_result" } ], "source": [ "cost(4)" ] }, { "cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "247.23339658094702" ] }, "execution_count": 28, "metadata": {}, "output_type": "execute_result" } ], "source": [ "cost(5)" ] }, { "cell_type": "code", "execution_count": 44, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "55.0" ] }, "execution_count": 44, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Example with Poisson demand\n", "stats.poisson(mu).ppf(a) # ppf is the inverse cdf \n", "# This gives a slightly different answer from complete enumeration when D has a discrete distribution.\n", "# There isn't a q such that cdf(q) = a.\n", "# As a result, our answer is off by 1 from complete enumeration. In practice, this usually doesn't matter, \n", "# unless we are only ordering 1 or 2 items." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Bisection Search" ] }, { "cell_type": "code", "execution_count": 45, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "55.0" ] }, "execution_count": 45, "metadata": {}, "output_type": "execute_result" } ], "source": [ "L = 0 # lower bound on qstar\n", "U = 1 # this will be an upper bound on qstar\n", "a = cu / (cu + co) # Newsvendor quantile\n", "\n", "\n", "while cdf(U) < a: # while the derivative of the cost is negative\n", " U *= 2\n", " n = n+1\n", "\n", "while U > L + 1:\n", " q = np.round((U + L)/2)\n", " if cdf(q) >= a: # the derivative of the cost is positive\n", " U = q\n", " else: # the derivative of the cost is negative\n", " L = q\n", " \n", "qstar = U\n", "qstar" ] }, { "cell_type": "code", "execution_count": 34, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "4.688232421875" ] }, "execution_count": 34, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Version for continuous demand\n", "\n", "L = 0 # lower bound on qstar\n", "U = 1 # this will be an upper bound on qstar\n", "a = cu / (cu + co) # Newsvendor quantile\n", "\n", "# Find U large enough that cdf(U) >= a\n", "while cdf(U) < a: # while the derivative of the cost is negative\n", " U *= 2\n", "\n", "#while U > L + 1:\n", "while U-L > .0001:\n", " #q = np.round((U + L)/2)\n", " q = (U+L)/2\n", " if cdf(q) >= a: # the derivative of the cost is positive\n", " U = q\n", " else: # the derivative of the cost is negative\n", " L = q\n", " \n", "qstar = U\n", "qstar" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Finding qstar using scipy.optimize (Derivative-free optimization)" ] }, { "cell_type": "code", "execution_count": 46, "metadata": {}, "outputs": [ { "data": { "text/plain": [ " fun: 90.7200000077868\n", " hess_inv: array([[0.03042296]])\n", " jac: array([0.07992744])\n", " message: 'Desired error not necessarily achieved due to precision loss.'\n", " nfev: 369\n", " nit: 4\n", " njev: 119\n", " status: 2\n", " success: False\n", " x: array([55.])" ] }, "execution_count": 46, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import scipy.optimize\n", "scipy.optimize.minimize(fun=cost, x0 = np.random.rand()*100)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Stochastic Gradient Descent" ] }, { "cell_type": "code", "execution_count": 80, "metadata": {}, "outputs": [], "source": [ "def stochastic_gradient(d,q):\n", " if q > d: # we ordered too many and had overage\n", " return co\n", " elif q < d: # we orded too few and had underage\n", " return -cu\n", " else: # q == d, which should happen with probability 0\n", " print('Q is equal to D! Derivative undefined')\n", " return (co + cu)/2" ] }, { "cell_type": "code", "execution_count": 106, "metadata": {}, "outputs": [], "source": [ "q = np.random.rand() # Choose a random starting point\n", "hist = pd.DataFrame(columns=['q','g', 'step'])\n", "for n in range(10000):\n", " d = demand.sample().values[0][0] # Choose a random demand from the data\n", " g = stochastic_gradient(d,q)\n", " step = (n+1)**-.5 # Any sequence that decreases slowly\n", " hist = hist.append({ 'q' :q, 'g' : g, 'step' : step} ,ignore_index=True)\n", " q = q - step * g" ] }, { "cell_type": "code", "execution_count": 107, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
qgstep
00.561783-5.01.000000
15.5617831.00.707107
24.8546771.00.577350
34.2773261.00.500000
43.7773261.00.447214
............
99957.6462201.00.010002
99967.6362181.00.010002
99977.6262171.00.010001
99987.6162161.00.010001
99997.6062151.00.010000
\n", "

10000 rows × 3 columns

\n", "
" ], "text/plain": [ " q g step\n", "0 0.561783 -5.0 1.000000\n", "1 5.561783 1.0 0.707107\n", "2 4.854677 1.0 0.577350\n", "3 4.277326 1.0 0.500000\n", "4 3.777326 1.0 0.447214\n", "... ... ... ...\n", "9995 7.646220 1.0 0.010002\n", "9996 7.636218 1.0 0.010002\n", "9997 7.626217 1.0 0.010001\n", "9998 7.616216 1.0 0.010001\n", "9999 7.606215 1.0 0.010000\n", "\n", "[10000 rows x 3 columns]" ] }, "execution_count": 107, "metadata": {}, "output_type": "execute_result" } ], "source": [ "hist" ] }, { "cell_type": "code", "execution_count": 104, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[]" ] }, "execution_count": 104, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAD4CAYAAADxeG0DAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAPcUlEQVR4nO3df5DcdX3H8dfrbu8CSS6SH0cSCEk4oFFIKT9WOOqMtQU7pGXA6tiK1jKtTOpMba3TjgrMqNMfM0611c7oaFNKsdOA06FQLBIFrK1/SBjvAOGQH9IriSEJXEIgIQm5u+y7f9wm5sclu3v7zW7e3PMxk+H2u7vf7/tuN0+++WQ364gQACCvjnYPAABoDiEHgOQIOQAkR8gBIDlCDgDJldpx0AULFsTy5cvbcWgASGtwcHBbRPQeub0tIV++fLkGBgbacWgASMv2hsm2s7QCAMkRcgBIjpADQHKEHACSI+QAkFwhIbd9m+2XbQ8VsT8AQP2Kevnh7ZK+IulfCtof0FaDG3Zo/fB29ffN16XL5tZ9u8nud8cjG7VuaItWrVysD16+9ODlCxbP0aMbd+hHL+wQ/wZpe1k67DHond2tV/aMqVIJzZvZJXdae/btV4elSkhzZ3Xr7cvmamjzTu0d2685M0rqLnXo7AWzNLR5p7btekMdtpbOm6mtO9/Q3JndunjZXK084y3asWe05vOq4fmL+mdsbS+XdF9ErKx123K5HLyOHCerwQ079KFb12t0vKLuUofW3tg/6W+6I2/3mWsu0F/c99Rh93t26y7dfM+TB+/znovO0H88vrmV3w5OMpY0o+vYz6vj3tcejIjykdtbtkZue7XtAdsDIyMjrTos0LD1w9s1Ol5RJaSx8YrWD2+v63brhrYcdb91Q1sOu89/P8dzf7oLHf95NRUtC3lErImIckSUe3uPeocpcNLo75uv7lKHOi11lTrU3ze/rtutWrn4qPutWrn4sPu86xd47k93HTr+82oq2vIWfeBkdumyuVp7Y3/NNfLJbrdiUc9hlw/c99A18svOns8a+UmGNfIDO2KNHABOqBO6Rm77TkkPS1phe5PtjxSxXwBAbYUsrUTE9UXsBwDQON7ZCQDJEXIASI6QA0ByhBwAkiPkAJAcIQeA5Ag5ACRHyAEgOUIOAMkRcgBIjpADQHKEHACSI+QAkBwhB4DkCDkAJEfIASA5Qg4AyRFyAEiOkANAcoQcAJIj5ACQHCEHgOQIOQAkR8gBILlCQm77atvP2n7e9qeL2CcAoD5Nh9x2p6SvSlol6XxJ19s+v9n9AgDqU8QZ+WWSno+I4YgYlfRNSdcVsF8AQB2KCPmZkn52yOVN1W2Hsb3a9oDtgZGRkQIOCwCQigm5J9kWR22IWBMR5Ygo9/b2FnBYAIBUTMg3STrrkMtLJG0uYL8AgDoUEfIfSTrP9tm2uyV9QNK3CtgvAKAOpWZ3EBHjtj8m6buSOiXdFhFPNT0ZAKAuTYdckiLifkn3F7EvAEBjeGcnACRHyAEgOUIOAMkRcgBIjpADQHKEHACSI+QAkBwhB4DkCDkAJEfIASA5Qg4AyRFyAEiOkANAcoQcAJIj5ACQHCEHgOQIOQAkR8gBIDlCDgDJEXIASI6QA0ByhBwAkiPkAJAcIQeA5JoKue33237KdsV2uaihAAD1a/aMfEjSeyX9oIBZAABTUGrmzhHxtCTZLmYaAEDDWrZGbnu17QHbAyMjI606LAC86dU8I7f9kKRFk1x1S0TcW++BImKNpDWSVC6Xo+4JAQDHVTPkEXFVKwYBAEwNLz8EgOSaffnhb9neJOkKSd+2/d1ixgIA1KvZV63cI+megmYBAEwBSysAkBwhB4DkCDkAJEfIASA5Qg4AyRFyAEiOkANAcoQcAJIj5ACQHCEHgOQIOQAkR8gBIDlCDgDJEXIASI6QA0ByhBwAkiPkAJAcIQeA5Ag5ACRHyAEgOUIOAMkRcgBIjpADQHKEHACSayrktr9g+xnbT9i+x/ZpRQ0GAKhPs2fkD0paGREXSnpO0k3NjwQAaERTIY+IByJivHpxvaQlzY8EAGhEkWvkfyBp3bGutL3a9oDtgZGRkQIPCwDTW6nWDWw/JGnRJFfdEhH3Vm9zi6RxSWuPtZ+IWCNpjSSVy+WY0rQAgKPUDHlEXHW8623fIOkaSVdGBIEGgBarGfLjsX21pE9J+pWI2FPMSACARjS7Rv4VST2SHrT9uO2vFzATAKABTZ2RR8S5RQ0CAJga3tkJAMkRcgBIjpADQHKEHACSI+QAkBwhB4DkCDkAJEfIASA5Qg4AyRFyAEiOkANAcoQcAJIj5ACQHCEHgOQIOQAkR8gBIDlCDgDJEXIASI6QA0ByhBwAkiPkAJAcIQeA5Ag5ACRHyAEguaZCbvsvbT9h+3HbD9g+o6jBAAD1afaM/AsRcWFEXCTpPkmfKWAmAEADmgp5ROw85OIsSdHcOACARpWa3YHtv5b0e5Jek/Srx7ndakmrJWnp0qXNHhYAUOWI459E235I0qJJrrolIu495HY3STolIj5b66DlcjkGBgYanRUApjXbgxFRPnJ7zTPyiLiqzmPcIenbkmqGHABQnGZftXLeIRevlfRMc+MAABrV7Br5522vkFSRtEHSR5sfCQDQiKZCHhHvK2oQAMDU8M5OAEiOkANAcoQcAJIj5ACQHCEHgOQIOQAkR8gBIDlCDgDJEXIASI6QA0ByhBwAkiPkAJAcIQeA5Ag5ACRHyAEgOUIOAMkRcgBIjpADQHKEHACSI+QAkBwhB4DkCDkAJEfIASA5Qg4AyRUSctt/bjtsLyhifwCA+jUdcttnSXq3pI3NjwMAaFQRZ+RfkvRJSVHAvgAADWoq5LavlfRiRPy4jtuutj1ge2BkZKSZwwIADlGqdQPbD0laNMlVt0i6WdKv13OgiFgjaY0klctlzt4BoCA1Qx4RV0223fYvSjpb0o9tS9ISSY/aviwithY6JQDgmGqG/Fgi4klJpx+4bPsFSeWI2FbAXACAOvE6cgBIbspn5EeKiOVF7QsAUD/OyAEgOUIOAMkRcgBIjpADQHKEHACSI+QAkBwhB4DkCDkAJEfIASA5Qg4AyRFyAEiOkANAcoQcAJIj5ACQHCEHgOQIOQAkR8gBIDlCDgDJEXIASI6QA0ByhBwAkiPkAJAcIQeA5Ag5ACTXVMhtf872i7Yfr/76jaIGAwDUp1TAPr4UEV8sYD81DW7YofXD29XfN1+XLpvbikPiJNfq50St49Vz/df/53/1k82vafe+ce0dq6i71KFzFszS8Lbd2je+X72zZ+iaC8/Qrn3jeu6lXdqxZ0zzZnZJkl6pfn3ewh6995IlenbrLn31+z/V3rGKfvvSJXr3BYu0fni7du0d08PD2zWjNHGu9uKreyVblf0Vvbp3TKUOa7wind4zQysW9WjHnlHt2DOmuTO7NHdmt0LSa3tG9cruUfX1ztYbY/v1xIuv6cIz36Il82bqh89v07bX9+nc3tl6fd+4tu0eVYekpfNmauvONzSj1KmuUofmzezSSzv3afvufbpk6Vxd3jf/4M/m7X/1oEZeH9Xs7k6dc/psjY5XNLa/onmzunXuwh7NmVHSXY9t0u43xrViYY9C0qZX9mjXvnGd3jNDe8Yq6rDU1WG9undMC+ecovMW9siSentmqGdG6eDP4MD+/vPJLdq1d0xvXdSjT616m57dukvrhrZo1crFkqR1Q1t0weI56jm167DHcHDDDt396CaFpPddsuSwx/aORzYe3McHL1/alk45IqZ+Z/tzkl5vNOTlcjkGBgYaOtbghh360K3rNTo+8cRfe2M/MZ/mWv2cqHW8eq7/nX/4ocYrxczT2WHtr8QR26RKRZr67+oT75SuDp1S6tCre8fbOod17J+TJc3omngMJen6f5x4XCWpu9O6c/UVunTZXN3xyEbdfM+TB+/30Xf26faHXzhhz0nbgxFRPnJ7EWvkH7P9hO3bbB9zYturbQ/YHhgZGWn4IOuHt2t0vKJKSGPjFa0f3t7U0Miv1c+JWser5/qiIi7pqIhPbDu5Iy5N/GzaHXHp+D+n0M8fw/XD2zV2yAM3tj8OPrbrhrYcdr/vPLW1LZ2qGXLbD9kemuTXdZK+JukcSRdJ2iLpb4+1n4hYExHliCj39vY2PGh/33x1lzrUaamr1KH+vvkN7wNvLq1+TtQ6Xj3Xlwp8eUFnhyfZNnE2eTLrKnXotFOLWNVtzvF+TtbPH8P+vvnqOuSB6+r0wcf2wJLMAVdfsKgtnWpqaeWwHdnLJd0XEStr3XYqSysSa+Q4GmvkrJFPpzXyYy2tNLtGvjgitlS//oSkyyPiA7XuN9WQA8B0dqyQN/vnm7+xfZEmlpRekPSHTe4PANCgpkIeER8uahAAwNTwzk4ASI6QA0ByhBwAkiPkAJBcYa8jb+ig9oikDS0/cDEWSNrW7iHagO97euH7Pjkti4ij3lHZlpBnZntgstdxvtnxfU8vfN+5sLQCAMkRcgBIjpA3bk27B2gTvu/phe87EdbIASA5zsgBIDlCDgDJEfI62T7L9vdtP237Kdsfb/dMrWK70/Zjtu9r9yytYvs023fZfqb6mF/R7plawfYnqs/vIdt32j6l3TOdKNVPNXvZ9tAh2+bZftD2T6v/TfHBB4S8fuOS/iwi3iapX9If2T6/zTO1ysclPd3uIVrs7yV9JyLeKumXNA2+f9tnSvoTSeXqB8R0Sqr5+QKJ3S7p6iO2fVrS9yLiPEnfq14+6RHyOkXEloh4tPr1Lk38xj6zvVOdeLaXSPpNSbe2e5ZWsT1H0jsl/ZMkRcRoRLza3qlapiTpVNslSTMlbW7zPCdMRPxA0itHbL5O0jeqX39D0ntaOtQUEfIpqH6s3cWSHmnvJC3xZUmflFTgxwaf9PokjUj65+qS0q22Z7V7qBMtIl6U9EVJGzXxGbyvRcQD7Z2q5RYe+NSz6n9Pb/M8dSHkDbI9W9K/S/rTiNjZ7nlOJNvXSHo5IgbbPUuLlSRdIulrEXGxpN1K8kfsZlTXg6+TdLakMyTNsv277Z0K9SDkDbDdpYmIr42Iu9s9Twu8Q9K1tl+Q9E1Jv2b7X9s7UktskrQpIg78iesuTYT9ze4qSf8XESMRMSbpbkm/3OaZWu0l24ulic8klvRym+epCyGvk21rYs306Yj4u3bP0woRcVNELImI5Zr4S6//iog3/RlaRGyV9DPbK6qbrpT0kzaO1CobJfXbnll9vl+pafCXvEf4lqQbql/fIOneNs5St2Y/fHk6eYekD0t60vbj1W03R8T9bZwJJ84fS1pru1vSsKTfb/M8J1xEPGL7LkmPauJVWo8p6VvW62H7TknvkrTA9iZJn5X0eUn/Zvsjmvgf2/vbN2H9eIs+ACTH0goAJEfIASA5Qg4AyRFyAEiOkANAcoQcAJIj5ACQ3P8Djk6A//SfyBIAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.plot(hist['q'],hist['g'],'.')" ] }, { "cell_type": "code", "execution_count": 108, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[]" ] }, "execution_count": 108, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAD4CAYAAAAaT9YAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3de3hU1bk/8O+aXAiBIEMSbsYEU1OKQgWSNkGRYpVa1FbwinjUY1XU1ueU1seWIrWn1sPBc1ofzvnpj4vaiz1EtCJoLWip1SIHE2GAlkukwZTEyC3gcL/kts4fs3eY2bP3nj0zezJ7Jd/P8/CQ2TOZWXPJO2uv9a53CSkliIhITb50N4CIiBLHIE5EpDAGcSIihTGIExEpjEGciEhhmd35YAUFBXLEiBHd+ZBERMoLBAKHpJSFZtd1axAfMWIENm3a1J0PSUSkPCFEo9V1HE4hIlJYzCAuhPilEOKgEGJ72LFBQoi1Qoh67X9/aptJRERmnPTEfw3g64ZjcwC8I6UsA/COdpmIiLpZzCAupVwH4DPD4RsA/Eb7+TcAprncLiIiciDRMfEhUsp9AKD9P9i9JhERkVMpn9gUQswSQmwSQmxqaWlJ9cMREfUqiQbxA0KIYQCg/X/Q6oZSyqVSygopZUVhoWmaoxICjUE8++5uBBqD6W4KEVGXRPPE3wBwN4AF2v+vu9YiD6qubcKPV21DpwSyMgRemjUB5SVMyCGi9HOSYvgSgA8AjBRCNAsh7kUoeE8RQtQDmKJd7pECjUHMW7kNHRKQAFo7JFZsbk53s4iIADjoiUspb7e46iqX2+I5gcYgnvj9DnQajr+1bR/mTx+TljYREYXr1mX3Klmwug6L1zWYXvfZqbZubg0RkTkuuzdRXdtkGcCJiLyEQdzEyxub0t0EIiJHGMRN9Mm0f1lEN7WDiCgWBnETx8+0217vz83qppYQEdljEDcRPNVqe/2gftnd1BIiIns9Ioi7vZpy2tjzba/PyugRLxsR9QDKpxgGGoO4fekHaOuQrq2mnHPtKNvslLYOY+Y4EVF6KN+lXLG5Ga0d0tXVlLF69G0dMunHICJyg/I98UPHz0ZcdiNzpKbhsO31jZ+dQqAxaNrjDzQGsWJzMwSAG8cXscYKEaWU0kE80BjEO3UHIo5dMvy8pO+3qjQ/5m0W/+VjPHdXRVR7bl60AXo/ffnGJrzywGVJB/JAYxA1DYdRVZrvuS+F6tomPPtuPYKn2jAoNwvfvrIMMyuL090sol5D6eGUFZubYRzZ2L73aNL3u2v/8Zi32dIUPZk6b+U2hDenozMU7JMRaAzi1iUb8J9v78KtSzZ4qhRudW0T5q7chk+PnMGp1g40HzmDuSu3obqWi6WIuovSQdw4lAIAuw/EDsCx/P9362M/9olW/PztXbh96QcINAYRaAziI5Pgv+WT5ILuU2vqoM+jdnSGLnvF03/aZXr8l+tZsoCouygdxAvz+kQd++ykfY63E04LXOmTqYv/8jFuX/oBzKY7T53tSKotf20+ans5nQ4dN3+td7ecxOT/fBcLVnvnC4eop1I6iOf1iR7Sbzh0Mukhh+yM+KZH/7LrIFotMlaSfYFb2yPTGVVJb9xz+BQWr2tgICdKMaWD+J8Mk5oA0CljZ5fE0h5nCqFVAAeAs0kE3UBjMKp3n+lTq3LLstpG0+MLVtext07kAmWyU8wyNA6diB4TB4Djp5Or932qLbkhkHDxfiGEW2IyKZrtkdWis5dvcXS742c7ItIxq2ub8G9/2ImTraHXWF9UNefaUalpKHlCdW0T1mzfh6mjhzF7yWXeiAgxBBqDuOP5Gvzij7twx/M1XcMlPote6ZJ1DQkPqSxYXYdOQ9xN14vU0HIi6tiJ1g7HATSV/rBtn+Pbzlu5DcC5bBY9gOte2fSJq20jb7nrhVrMXbkN79cfYvZSCigRxGsaDqO1vROdEmhr7+waLhltkRMugYRXbpqd/udkZyR0X0DssrZ2si1+d9XWvfjyv61N6x9Dp3R+hqFn7VhlrRjH/alnqK5twuUL3sG6+kMRx3/2+x1palHPpMRwSlVpPrIzfWhr70SGT2DvkdMINAax98hpy98xSz904qRJNkky33RDBuQk/Lt2hbYOHm/FXK2Hm47T06KBuWj87JSj20qE/qB3t5w0vb7NeOpDyrPb3vC0y1/as5dvwXt/b8Hkzxdi4Yxxrt63CpQI4uUlfjx+/SV4eWMTdu47hpc+bMKKzc0osCkJa5Z+6ITZx0v4BPpnZ+BEa/xj5e2diX9gJ5Tmx0wp/OX6hrQE8dw4z070LxwzybxG5D2zl2/Bqq17u/2xVm3di7U7D6C8xA8J9JrxdyWGUwKNQTzx5g78rfko2jokOiXQ2tZpO2noxvJ7XdHAvph73cUJ/W4y2SnHz9pvTgEAnzjsDbut4ZB5rzoRHZ3AiDl/QNljqzleqrgFq+scBfBpz6x35fHeMDzWydYOrKs/1KvG35UI4vqYeHjI7gSw32bIZNWWxMbEzVLEx5f4MbOyGLnZ8b9cfTITH0//u4PVp2dTUFHRSfpfewry1ds6ZK/5w+spFqyuQ9X8P+HWxaGSEM+972y17tbmo0lN0Ot7CMT6FP54lfUZYE+hRBDXx8TjyZDevvdYYg8moh/lxvFFAICywXlx312myf05tePT7l+dqY9l6ot1pvziPYx94o9Rf3DGeU03P0hPrzVfzk/eon9W9h87iw/3BHHTog1RtYzsvPm3xIZcAo1B3LIoVE8olg4ZOsub8ov3EnosFSgRxMtL/Fh2XxUqRjiv4NeR4GRZtiFtMTtDdOU43/al+MfX9LK1iTjd1v1jxUsNPan6lpM4cqoNq7bujQjkxpa52dJDJ5IvnUCpt8Rm4xQn2juBO1+ojfvMa97KbXF/3upbTvbYQK5EENe1xJFx0jcrsaeWmRkZxPuE3c/MymIkkjGYaCVDp4sz3Rx+sPvu+/1f99o+3rSxw11rB3WPZFbOujGQp49dj/3p245/x0mVUTP1FtlRqlMiiOuLffYcdj6JV5DnPLUvfI/OooG5EdcZLyfSwf/wH+fKADjdD7S6tsnxqencldtcKVEb68ugQ4bab5bvPXRAHyycMQ7Txg7HwNyspNtCqWccOktnCYQjp9vxhXlrHH2Okznr88JCObcpEcT1ic14jB4+wNHt9D069bKyeTmRWZcXDIoM4na9YwHgirKCqLOAo6fbMXv5FgQag7hNqw1+W4za4FZlXq0scKFE7b++sd3R4xw7E13W4Nk7ygEAC2eMw9bHv4aigYnnxwPosae+XmKchIxneCQVk89n2jtx06JzfxeBxiDmrtyGx1zqpACJj8N7mRJBXJ/YjMd2h5OCxj066/ZHTogWGPLNpc30ar/sDPz23kpkmExmrtq6F/NWboP+XdQeY8MIqzKvVnYkOpEbxq6Ql+6jfceiFiEV9M+O2nHo21eWJdWWnnrq6yXGt1tflOXEYzZ5/8mapwXtmxZtQHVtE5bVNkUE92T0xMXBSgRxfbHPiPzc2DfWOF0VtsXwwTh+5tyCnswMgZu0zJSuYzZdcX3lodWioF2GlMGdce5ClOkTeHBSqfljJ5ny5/SPV8roSePSgn5Rt+sNiyx6oh+v2ob7X9yEG55Zb/uZSOUa27r9x3H70g+ijs9YsqFHDockS4kgri/2iWdMfIBJrXEzdkvHx10wMKqHOdRmGX1GjJlI43j6UYebT+iyM32W1f5kHLVMzPzHW86GY060dmD/scgJ5t0Howt1AfZ1Y5ycWLFMbYjTeZR4WC226ZDA2p0H8Nfmo6Y5+9W1TbjzhdqY93/R4P6YP30MVjx0GaZcPASXFp0X18S32VlhWyccLSQalJuFPQuuQ1lhdOeiJ1Ji2X0iY+LHHKx2BOx7sPq2a+GB3G6JeGH/+Jb6G6v5hfOJ6KCvL0QSiO4JJXuaeOS0s9fLjNVOSPdcNsK0fsaksgJ89+rP46ZFG2zvd1ltY68vUasPK+hWPJT8xtvVtU3Y6nCHqHkrt6Hp8Em8tWM/igflRhWzsvKn73+l6+fwDcW7Yzn+c3d/CQCw9pHJGDHnD3H/vmplc5UI4vqYeGtbp+OZ6ZMmk29m7PrO+gYT4X80dl8m8dYAkUBX78pYKz0vJxNHDYE1S+u+FvTPRosCudR6AH5rx36MvWAgyobkRTzHaWOH2/5Rn+mJA5gOhNfOv9nwRXfLog1oWHBdwvddXduEx193Pp7diXM13+M5E7ZSmILP7vzpYzByaF7U35CVac+sx6qHJ5peF16L5f36Q3hr+z68eG+lq+11mxJBvLzEj3+eMAJv7diPM23Rp/Nmjpxuj+pFx0sg9AUSzm7Qom926OWMp1jWgjV1COwJolN7vJmVxbhxfBFGDsnDh3siT59vLb8AAPC9KSNtC0p5yZxrR1n2phfOGIehA3Lw1o79GNg3K6p3mMyGGqrSVyNafX11Arh18Qb8cOqomJ/t8B4lADz1Vl1UxyAV7DKTNs6bgi89uRYtJ1pR2D8bi++siHlGFoveWzZ7PQbkZOLYmcjnbHUWUl3bFNWpWFd/CAtW10V9hhesrsNbO/bj65cM7brO7Fh3SCqICyG+B+A+hGLbNgD3SCnPuNGwcNW1TZZlLe2s2Nwc84NuFybyTbIu+mRZ10L51uUXAgDmXnex4yC7uSnY9QcrASyrbcLyD5swdGDfiNuVDMrt+mDMrCz2VBDP65N4fZjwIG889e19IdzZakR9ibvV0Mrs5Vvw9o79XSt+33c4BOKGooE5WD/nKtvbbJw3JeJypk+gPcEV1v1iVNP81T1fdvwlYZViu3hdQ0RQDi+zu3hdA96pO4CrRg2JOLZz37Fu68EnPLEphDgfwL8AqJBSjgaQAWCGWw0Lt2a7811kwu12UEDKZ1PbxCxgXzIsMv981NA8XFFWgPnTx3T1COIZRzMbku+QwKfByFrpgwdEjrfnmMwM6kWIzFTXNuGGZ9Zj1oubom6TbE72r7+Vug9rbyuGVRfHasTv/E8g4nJ1bRMqnlyLVVv3dnvJhvnTx2DPgutiBnAz139xWMKP+1iM6qLxnIk7SbEFgBdrIjeOqW85iaWGTua6+kOorm3qlr1kkx1OyQTQVwjRBiAXQEpmLaaOHpZQb+LTo7FPCuy+xTpMxmQf+Mrn8O6ug2jrkMjKEHhy+hjTD0qGiM7DTcZew3P5YtF5UcMtH+4J4tYlG/DKA5E9NH1btJCj+OPOA+jfJwNXjxoCwDwn22xi1UxudkbSE212nn23XonJpXQIr+IZ+R53v2Teo4UzxiU04Tlt7PBu+2yED82eNhkqNfvK/Nc3tnd9MaRyL9mEe+JSyk8B/BxAE4B9AI5KKf9ovJ0QYpYQYpMQYlNLS0tCjzVyaF7M9D0zrQ42PG63Sc07dDJ6Aqa8xI/lsybg0WtGYvmsCZYBbNh5ya1YjGJo5w+nmn8YOkwWET1lkj544mwHVm3di9ct/nhKC/s7apbfxSX2WSZ1gD89ciahIkmqirc2z6h5awCoX/kx3j/vByeVOt7Fx6y89F0maZJ2L334kIzTvpmxZ/+cxfaEyUpmOMUP4AYAFwIYDqCfEOKfjLeTUi6VUlZIKSsKCwsTeqzXNjcnVJXwVGtHzNxau93jrcbpykv8+M6VF9n2QJNdsWg0ICcyWJaX+E0/nADwwe7Isxa7ySyrV/Vbl19oGlSNvuPi87xujPlpdU8v8D97+Zaucr9XfmFIXL97ur0TgcZgWis/DshJfE5E981LzXPIc0wK2T04qTSuHu03TO7bLFUy1m5Vs5dvcZQjbyUFJfgBJLfY52oA/5BStkgp2wC8BuAyd5oV6WCi+2W2duDWGDVKhhkmEN0ys7LY1ap+ZvnsVsM1Jxx8ecUys7IYy2dNwIAc+xE3N09nY/Ws5v9hp2uP5QXhY9h6ud+WY/HnBdz/m41JtcPsu9rB93eXX92T/JzIwhnjMKmsIOr449dfgmljh6Nvlg+D87Ixf/qYuIcknPbYY+31umrr3m6dJHYqmSDeBKBKCJErhBAArgKQktH7wQnulwmEvv3m2YwVXv2FwQnfdywLZ4zD/OljXLkvp8MbOjcKYpWX+DHHYtgmHRLZ49SrFqyuw9yV26J60Gbpb7EyMKwWWzll1hn4xqXDTYOqkRuLj3Qv3lsZscJTTxZYOGMc6n42FR8+NiVlY+CBxiDOdsO6hFSUDUhmTLwWwKsANiOUXugDsNSldkVwul+mVa+xbv/xrtWXxuXLdis74+mNWJlZWYySQc5rvlh54Cufi+v2G/cEXVmmPbOyGPOnj8EVDv6gyZlAY9BxyqxPADue+Lorn0U7Kx66DKOG5qF/nwxMGzscC2eMw4v3VuLBSaUYkZ9ruoR9UlmB65Pa5SV+PHdXBV5/eGLKJy3Dh+eecqHT40QqVqwmlZ0ipfwJgJ+41BZLwVOtpkvNjeyKUy1YU4ctnxxBe4dEZobAy9qk5CGboRq3PqBP3zY2qQUNIsG2zFu5DSOHxr+lnNHMymLMrCxOaAlzvMYWnRdzSXg8y6LTtQDDzpI4NgnRz/Dvv6I0obUSurFF56FwQA7W7jxgen15iR9rZk+KOh6exz97+Rb8/q97IQFMvKjA8ysZw/XLzogqczFX+/soL/HjowQ3mvACJVZs+nOzHc0IDxmQY3lquWlPsOs+2jsknlpThx9OHYX3dh20vD+rDJB4lZf4seKhy7BiczNWBJrjPm2zmmDMyfTZLk2v2388oaXSX7vYfHKtZFBuRMEwN84wjFY9PNH2yyL8On180iqQGxdlAKlJ8YrX+vr4s7T0dq/a+qmjFcvhwtcwmKUiOi0UtXDGOMfjy15zZ1WJ6Zfg91/eir/84ErkmazsVIUSVQyDp2LPvA/KzYrawCGc8Uvgwz1BvLa5GW0Ws4OjtG9ot5SX+DF/+hjk9Y3/ezPTIoPmi0Wxh5lOO0izDCdgPXTz9G1ju2rNCO1yKmTFkW9mVwfEuMmB053YU+1UHAtxBoZ9XuZcOwo1c692/Lvh48o6fXgsR9t4vKywH9Y+MtnxfarK6su78bNTqK5twmHD3ERh/+y4Sl+nkzI98Vj6xpj8MfP3A8eRYbHktygFvUwA6JMZfzuzMs2D2g+njkq67oRRhs966Ka8xI9XH7rMcaGhRP30htGOF660d5oPr1TXNkV9catYiuUHX48OPsYzIjN2E4768Fhv0zfTZ7rPgNln7dCJVkd1XZwM/4VLxdSGMj3xWE++r01NEyv1B47jvokXml6XTEaMnfMTWARU2N/8d8pL/Elvg2YUqyy5kxz5ZM2sLEZ+HIuI5q7c1pVLvmB1Haprm/CzN3ekrH1O6HW3k81tNwu2Ts6AUvn+qKruyamObytxbhh01NC8qECZpW3QsurhiXhwUimyfIAQoaAefqZjnKcrSUHvXpmeeKxO1Ojzz8OpOFPQjpxux3Pr/2F63Y2GHX3cUmZSnTAWvbCWmW9fWWbZa3UyGRz9SylOg3Bo6d1fws2LNsTd/mQm/9wSPu5sHLd3I2MoVoDun8BZKZkLn/ANLxEc/h6YVeoMf79vWbwBnTKUafSLW90fglSmJx6L0z01jaxWgqaqJxPvl8OXRvhtT31nVhZb5vNKxJ8mOcgjO9XrQzdOVo3GY/Tjb3X9YVXN/1NKChMZFyWFX35tc7Pj+8k1Wa3oxNwYRaF6s2QCXiJnoeUlfvzuwcvw6DUj8bsH3cupD6dEEHcyJn4mzgk8O6kMZOUlfsv6GMb5PJ+Ao8U2ej6vWbyLdxz4e1NGxvcLKVRe4sdPvzna1fs80dqBmxZtwMY9Qew/dhaL1zW4HsiNi5LCL8fzdkxL4Gywf3ZGrxzvdqogL3YsAULDIm5J9RCkEkE8eKo1ZoGci4efh0KXxrGnWtTwcEs/i0VJ2Rk+PDipFEMH9MGXR/jj+uaec+0oXDUqdt2NUTZ5430yhecCwMzK4pTvlbistjH2jWKorm3C5x9bbZkeqX9RnDJZXGZVnsG4SXc4q9eEvXB7s6+O3UnJ9MFy5x8vUiKIV5Xmwxcjik8eOdjxyk472Zm+lI2H66wmKvP7Z3elkb2SwKlXrFWdGT7gyeljcNFg8yX81fdPiOvxusvaRyZjUllBylYtJrsNnD4GblePWh+r//NHkesSMnyh/GvjptJZGcL2/V/7yOSI9EMgtILSa1/CXhPr9SkamIPd8xPf/i4dlAjiu/Yfj7lV1/a9R7F9b2Lj4roMn8BL91elfGbfaqIy2YqAsdo9ICcT5SV+y8f3ckbDi/dW4uN/vw5fHuF+GztjpeTE8O+rnRfmOmnoiespp9X3V0UcXz4r9hfq1p9c07Us/sFJpUqtoEwnq6GSaWOHJ7SpRbopEcSd7OwjANsl9M7Ibglk+oKLiwr7oSAvGxcN7h+1KCNRfWy6q8FT7V2Pryq3VtGGSyaGV9c24fhZZ/Mx055ZHzVHob9dejrbo9eMjKuo1JxrR+G9R6/0xEpUVax6eGJUINfrxahIiRTDWDv7ZPgEbhxflHQRm2xf932npWrBhfA521JoUllBRE1lJxXrvEBf+ermLjYJ7DfS5ZdxFPo3WxRyIuwLoLzE7+mzoZ5EpTHvWJToieu1ua06mfdPvBDlJX5H27HZ6UjytNoLzjpc0v3ivZWYVFaAnCwfJpWpVcxoZmWxq9kDyRTr322ytV081P/EUbop0ROvrm2yLeGY1zeUEiiSDMJON0r1MqsyAmZUCtxGsQpl6Qr7Z6Mlxq43ib7rZlt8EXU3JXridmPiPhHKXgGA4SnapUclQ2yW9TvNkVWFk0nOjfOmdNVDnzZ2OC61mtR6Zn3cj2+2xRdRd1MiiF8ybIDldTlZvq5xxIEOFgXZUeLFiOE7ky+yvO77DnJkVfLDqaO6NtDO8ImolLsHJ5UCCA2//PbeSiycMQ6vPzzRNMd6a/NR13rW/bMzcEVZAQr7x/48ur0ilXofJYZT9OESM5kuTkZmx7vVuAfpk6V6Vb/wn1XOSjFTXuLHKw9MiKhn4WQTiLWPTDYdillXfwjVtU2OXie7Gihzr7u46z5iDfmMu2BgzMcisqNEELdbdl9acK5XleyKzfwUVS7sbsbMl54WvMMZMzrMihHF46m36xy9XjUNh02Px5MqmuETKUmZpN5Fia6n3bL7I6fP7eRz4/iipOr12g1FUO9w9FS7o0qDv/7f6OqX08YOjwrgdp/HVx6YwJRCSpoSQdxu2f3AsKGW8hI/PmexpDwWtxbbkBrscsNvWbQhZiA3y3gxWyxitcF0ovumEhkpEcTtlt2H98SByOEVM/pkV7icTB8DeC/zzUvNi04BQCdCm0y74cV7K0174zdYFL0iipcSQdwuxbDYsI3aA1/5HDItZvwzfaExU2PWQE8ZCyfnFs4YB7ty3bF2PzfOgdvNiT9g6DiUFfZTdok3eY8SQVzPsjBj7J+Xl/jx8qwJePSakZZ/WMaa2RwL751+esMYy+vsFgAFGoNR9VaGDLDOz59z7aiIQlW9YWNi6j5KZKfMrCzGkr98bLo5rFkOuZ6x8P/eqUd757k11e2doT9AYxoeh1J6p5mVxZi3chviWXUfaAyabhsXK9U12awZIitK9MQDjUE0B81397bLIc8z2XxBTw3TF4AwgPduWTbjIFN+8V7UsZqGw6a99L1H7HefJ0oVJYJ4TcNh08J8WRmia8m9mXHFkbP/4Uv0iQBg6uihltfVt5yMylKpP2A+Vu6w7hiR65QI4lWl+aYNvaXiAts0rfBJTp8Anpw2hmldFCHWBOP9v9kYcXldfYvp7TKU+EuinkiJMXEAocTasN64T9jvQQicm+QMX5ZNZOQDLMfFPzsVSmFdsLoOz73fYFmq/RtfZMogpYcS/YeahsMwVlcVwtnazFTvNE3qe3K6dZYKEArgi9dZB3Agdo+eKFWUCOJmwykdndKyfgVRPPTt8s4faJ4mqG9ybMXNDSqI4qVEEAdgWoTCrjAWUTxmVhbjfxPcJLcnbfVF6lEiiJsNpwChwlhEbhpgkpZK5GVKBPGq0nzTgkXsiZPbZn6Z6wZILUoE8fISPyZeFF0Njj1xclu8qyq5MQ+lW1JBXAgxUAjxqhDiIyFEnRBiglsNCxdoDGL97sj9DLlwh7xgmM2epkTdIdkBwP8C8JaU8mYhRDaA3Fi/kIjXNjdHjYkPOy+HaYOUdt++sizdTaBeLuGeuBBiAIBJAF4AACllq5TyiFsN0wUag/jdpk+ijl8ynGldlBrzbfLGB/bNxPzpY3BFWQE3EiFPSKYnXgqgBcCvhBCXAggA+K6U8mT4jYQQswDMAoDi4vg/8DUNh9FmssriTFtHAk0mim1mZTHmWmwKsfUn13TdhsgLkhkTzwQwHsAiKeU4ACcBzDHeSEq5VEpZIaWsKCwsjPtBqkrzkWGSmvK3T4/G32Iih8yKG3IOk7womSDeDKBZSlmrXX4VoaDuqvISP564YXTU8cmfj/8Lgcip+yZGb+PHLdXIixIeTpFS7hdCfCKEGCml3AXgKgA73WuavaE2O6kQJUtPNXzxgz3okKGStayPQl4kpHGfqXh+WYixAJ4HkA2gAcA9UkrLbcIrKirkpk2b4nqMQGMQty35AO2G9JQR+bl479Er4280EZFihBABKWWF2XVJpRhKKbcCML1jt9Q0HEaHyZp74wbJRES9kedXbFaV5iPLZFnc0dNtaWgNEZG3eD6Il5f4MXnk4KjjgzkmTkTk/SAeaAzivV0Ho473y85IQ2uIiLzF80G8puFw1KQmAGz9xPXFoUREyvF8EK8qzUemyWKfsRcMTENriIi8xfNBvLzEj1sqLog63q8Pi/cTEXk+iAPAjeOLohqaeHY7EVHPoUQQBwBhaOloVjEkIlIjiNc0HEZnZ+Qx7upDRKRIEDfbY5P7axIRKRLEAQDiXBQXYE+ciAhQJIiHhlPOTWVyf00iohAlgrixforPJG+ciKg3UiKIG3PFOzokahoOp7FFRETeoEQQDzQGcfD42a7LneDEJhERkGQ98e4QaAzijudrcLbtXI6hT3Bik4gIUKAnXtNwGK3tnV0rNAWA7EwfJzaJiKBAT9yfm43wIoZTLh6CBwFbGKEAAAoTSURBVL7yOZSX+NPXKCIij/B8T3z73qMRlwvy+jCAExFpPB/EjcmETC4kIjrH80H8xvFFyNRyxDMzBG4cX5TmFhEReYfngzgA+ISA0P4nIqJzPB/EaxoOo70jlJ3S0dHJRT5ERGE8H8SrSvORnelDhgCymFpIRBTB8ymG5SV+LLuvCjUNh1FVms/MFCKiMJ7viRMRkTXP98T1Zfet7Z3IzvRh2X1V7I0TEWk83xPXl913SqCtnRObREThPB/EObFJRGTN88Mp5SV+PH79JVizfR+mjh7GoRQiojCeD+KBxiCeeHMHWts7sXHPZxg5NI+BnIhI4/nhFI6JExFZ83wQryrNR6YvtOw+wyc4Jk5EFCbpIC6EyBBCbBFCvOlGgyweJPJ/IiIC4E5P/LsA6ly4H1OsnUJEZC2pIC6EKAJwHYDn3WlONKYYEhFZSzY7ZSGAHwDIc6Etplg7hYjIWsJBXAhxPYCDUsqAEGKyze1mAZgFAMXFxQk9VnmJn8GbiMhEMsMplwP4phBiD4DlAL4qhPgf442klEullBVSyorCwsIkHo6IiIwSDuJSyh9JKYuklCMAzADwZynlP7nWsjDVtU2484VaVNc2peLuiYiU5fkVm9W1TZi7chsA4P36QwCAmZWJDcsQEfU0riz2kVK+J6W83o37MlqzfZ/tZSKi3szzKzanjh5me5mIqDfz/HCKPnSiVzHkUAoR0TmeD+JAKJAzeBMRRfP8cAoQKkf77Lu7EWgMprspRESe4vmeOPfYJCKy5vmeOOuJExFZ83wQZwEsIiJrnh9O4R6bRETWPB/EuccmEZE1zw+ncEyciMia54M4x8SJiKx5fjiFm0IQEVnzfBAHuCkEEZEVzw+nEBGRNQZxIiKFMYgTESlMiSDOAlhEROY8P7HJAlhERNY83xPnYh8iImueD+Jc7ENEZM3zwylc7ENEZM3zQRzgYh8iIiueH04hIiJrDOJERApjECciUhiDOBGRwhjEiYgUxiBORKQwBnEiIoUxiBMRKYxBnIhIYQziREQKYxAnIlIYgzgRkcISDuJCiAuEEO8KIeqEEDuEEN91s2FERBRbMlUM2wE8IqXcLITIAxAQQqyVUu50qW1ERBRDwj1xKeU+KeVm7efjAOoAnO9Ww4iIKDZXxsSFECMAjANQa3LdLCHEJiHEppaWFjcejoiINEkHcSFEfwArAMyWUh4zXi+lXCqlrJBSVhQWFib7cEREFCapIC6EyEIogC+TUr7mTpOIiMipZLJTBIAXANRJKZ92r0lERORUMj3xywHcCeCrQoit2r9rXWoXERE5kHCKoZRyPQDhYluIiChOXLFJRKQwBnEiIoUxiBMRKYxBnIhIYQziREQKYxAnIlIYgzgRkcIYxImIFMYgTkSkMAZxIiKFMYgTESmMQZyISGEM4kRECmMQJyJSGIM4EZHCGMSJiBTGIE5EpDAGcSIihTGIExEpjEGciEhhDOJERApjECciUhiDOBGRwpQI4oHGIJ59dzcCjcF0N4WIyFMy092AWAKNQdzxfA1a2zuRnenDsvuqUF7iT3eziIg8wfM98ZqGw2ht70SnBNraO1HTcDjdTSIi8gzPB/Gq0nxkZ/qQIYCsTB+qSvPT3SQiIs/w/HBKeYkfy+6rQk3DYVSV5nMohYgojOeDOBAK5AzeRETRPD+cQkRE1hjEiYgUxiBORKQwBnEiIoUxiBMRKYxBnIhIYUJK2X0PJkQLgMYEf70AwCEXm6MCPufegc+550v2+ZZIKQvNrujWIJ4MIcQmKWVFutvRnficewc+554vlc+XwylERApjECciUphKQXxpuhuQBnzOvQOfc8+XsuerzJg4ERFFU6knTkREBgziREQKUyKICyG+LoTYJYTYLYSYk+72JEoIcYEQ4l0hRJ0QYocQ4rva8UFCiLVCiHrtf792XAgh/lt73n8TQowPu6+7tdvXCyHuTtdzckoIkSGE2CKEeFO7fKEQolZr/8tCiGzteB/t8m7t+hFh9/Ej7fguIcQ16XkmzgghBgohXhVCfKS93xN6+vsshPie9rneLoR4SQiR09PeZyHEL4UQB4UQ28OOufa+CiHKhRDbtN/5byGEiNkoKaWn/wHIAPAxgFIA2QD+CuDidLcrwecyDMB47ec8AH8HcDGA/wAwRzs+B8BT2s/XAlgDQACoAlCrHR8EoEH736/97E/384vx3L8PoBrAm9rlVwDM0H5eDOAh7edvA1is/TwDwMvazxdr730fABdqn4mMdD8vm+f7GwD3aT9nAxjYk99nAOcD+AeAvmHv7z/3tPcZwCQA4wFsDzvm2vsK4EMAE7TfWQNgasw2pftFcfCiTQDwdtjlHwH4Ubrb5dJzex3AFAC7AAzTjg0DsEv7eQmA28Nuv0u7/nYAS8KOR9zOa/8AFAF4B8BXAbypfUAPAcg0vscA3gYwQfs5U7udML7v4bfz2j8AA7SAJgzHe+z7rAXxT7TAlKm9z9f0xPcZwAhDEHflfdWu+yjseMTtrP6pMJyifzh0zdoxpWmnj+MA1AIYIqXcBwDa/4O1m1k9d9Vek4UAfgCgU7ucD+CIlLJduxze/q7npl1/VLu9Ss+5FEALgF9pQ0jPCyH6oQe/z1LKTwH8HEATgH0IvW8B9Oz3WefW+3q+9rPxuC0VgrjZmJDSeZFCiP4AVgCYLaU8ZndTk2PS5rjnCCGuB3BQShkIP2xyUxnjOmWeM0I9y/EAFkkpxwE4idBpthXln7M2DnwDQkMgwwH0AzDV5KY96X2OJd7nmNBzVyGINwO4IOxyEYC9aWpL0oQQWQgF8GVSyte0wweEEMO064cBOKgdt3ruKr0mlwP4phBiD4DlCA2pLAQwUAih7/Ea3v6u56Zdfx6Az6DWc24G0CylrNUuv4pQUO/J7/PVAP4hpWyRUrYBeA3AZejZ77POrfe1WfvZeNyWCkF8I4AybZY7G6FJkDfS3KaEaDPNLwCok1I+HXbVGwD0Geq7ERor14/fpc1yVwE4qp2uvQ3ga0IIv9YD+pp2zHOklD+SUhZJKUcg9N79WUp5B4B3Adys3cz4nPXX4mbt9lI7PkPLargQQBlCk0CeI6XcD+ATIcRI7dBVAHaiB7/PCA2jVAkhcrXPuf6ce+z7HMaV91W77rgQokp7De8Kuy9r6Z4kcDiRcC1CmRwfA3gs3e1J4nlMROj06G8Atmr/rkVoLPAdAPXa/4O02wsAz2rPexuAirD7+haA3dq/e9L93Bw+/8k4l51SitAf524AvwPQRzueo13erV1fGvb7j2mvxS44mLVP83MdC2CT9l6vQigLoUe/zwB+CuAjANsB/BahDJMe9T4DeAmhMf82hHrO97r5vgKo0F6/jwE8A8PkuNk/LrsnIlKYCsMpRERkgUGciEhhDOJERApjECciUhiDOBGRwhjEiYgUxiBORKSw/wOWFS0QmHHBsgAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.plot(hist['q'],'.')" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.7.4" } }, "nbformat": 4, "nbformat_minor": 2 }