{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Programmering i Python - Halveringsmetoden\n",
    "\n",
    "Målet i denne økta er å kunna løysa likningar numerisk vha. eigenprogrammerte funksjonar.\n",
    "\n",
    "I forrige økt såg me på bruken av variablar og tilordningar.\n",
    "Dette er ein av dei to grunnleggjande mekanismane i imperativ programmering.\n",
    "Den andre grunnleggjande mekanismen som me må kunna er kontrollflyt.\n",
    "\n",
    "## Kontrollflyt\n",
    "\n",
    "Når ei kodeline vert køyrd i programmet, seier me at lina har *kontrollen*. \n",
    "Det er ikkje alltid at me ynskjer at kontrollen skal flyta vidare til neste kodeline.\n",
    "Av og til vil me hoppa over kode avhengig av verdien på ein eller annan variable.\n",
    "\n",
    "Lat oss definera ein variabel til testen."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "testvar = 10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Vilkår\n",
    "\n",
    "Alle imperative programmeringsspråk har ein vilkårssats.\n",
    "I python ser han slik ut."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Positivt tal\n"
     ]
    }
   ],
   "source": [
    "if testvar <= 0: print( \"Ikkje-positivt tal\")\n",
    "else: print (\"Positivt tal\") "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evt. kan ein ha meir enn to alternativ."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Positivt tal\n"
     ]
    }
   ],
   "source": [
    "if testvar < 0: print( \"Negativt tal\")\n",
    "elif testvar > 0: print (\"Positivt tal\") \n",
    "else: print( testvar )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Me kan bruka vilkårssatsen inni ein funksjon:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def testfun(testvar):\n",
    "    if testvar < 0: print( testvar, \"Negativt tal\")\n",
    "    elif testvar > 0: print (testvar, \"Positivt tal\") \n",
    "    else: print( testvar )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-17 Negativt tal\n",
      "3.4 Positivt tal\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "testfun(-17)\n",
    "testfun(3.4)\n",
    "testfun(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Løkke (repetisjon)\n",
    "\n",
    "Alle imperative språk har òg løkker.  Der finst gjerne fleire variantar.\n",
    "Den mest kompakte er for-løkka, som ser slik ut."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Repetisjon nr.  0\n",
      "Repetisjon nr.  1\n",
      "Repetisjon nr.  2\n",
      "Repetisjon nr.  3\n",
      "Repetisjon nr.  4\n",
      "Repetisjon nr.  5\n",
      "Repetisjon nr.  6\n"
     ]
    }
   ],
   "source": [
    "for i in range(7):\n",
    "    print( \"Repetisjon nr. \", i)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I for-løkka er der alltid ein indeksvariabel (`i` i dette tilfellet) som vert definert som ein del av sjølve løkkesyntaksen. I `while`-løkka er det ikkje tilfellet:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Repetisjon nr.  10\n",
      "Repetisjon nr.  8\n",
      "Repetisjon nr.  6\n",
      "Repetisjon nr.  4\n",
      "Repetisjon nr.  2\n"
     ]
    }
   ],
   "source": [
    "while testvar > 0:\n",
    "        print( \"Repetisjon nr. \", testvar)\n",
    "        testvar = testvar - 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Oppgåve\n",
    "\n",
    "Merk at ``testvar`` har endra verdi.  Kva skjer om du køyrer dei blokkane over\n",
    "(``if``- og ``while``-satsane) over ein gong til?\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Eit enkelt døme\n",
    "\n",
    "Lat oss sjå på tredjegradsfunksjonen $f(x) = 3x^3 + 10x^2 - 50x - 3$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "def f(x): return 3*x**3 + 10*x**2 - 50*x - 3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Me kan testa funksjonen ved å plott:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7ff18423beb8>]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "plt.plot()\n",
    "x = [ i/20 for i in range(-140,90) ]\n",
    "y = [f(i) for i in x]\n",
    "plt.plot(x,y,\"-\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Me ser at funksjonen byter forteikn mellom 0 og 1, og sidan han er kontinuerleg, må der vera eit nullpunkt der. Me kan bruka halveringsmetoden for å finna nullpunktet."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.05859375\n"
     ]
    }
   ],
   "source": [
    "def bisect(f,lower,upper):\n",
    "    while abs(lower-upper) > 0.01: \n",
    "        midpoint = (lower+upper)/2\n",
    "        if f(lower)*f(midpoint) < 0: upper = midpoint\n",
    "        elif f(midpoint)*f(upper) < 0: lower = midpoint\n",
    "    return (lower+upper)/2\n",
    "print ( bisect(f,-2,2))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Oppgåve\n",
    "\n",
    "Finn dei to andre nullpunkta vha. halveringsmetoden."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Forenkling av funksjonen\n",
    "\n",
    "Legg merke til at ``bisect`` reknar ut ``f(midpoint)`` to gongar.\n",
    "Her er det ikkje noko problem, men dersom det tek tid å rekna ut $f$ er det bortkasta arbeid.\n",
    "Det kan difor løna seg å endra funksjonen slik:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.05859375\n"
     ]
    }
   ],
   "source": [
    "def bisect2(f,lower,upper):\n",
    "    while abs(lower-upper) > 0.01: \n",
    "        midpoint = (lower+upper)/2\n",
    "        fm = f(midpoint)\n",
    "        if f(lower)*fm < 0: upper = midpoint\n",
    "        elif fm*f(upper) < 0: lower = midpoint\n",
    "    return (lower+upper)/2\n",
    "print(bisect2(f,-2,2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Rekursjon\n",
    "\n",
    "Hittil har me brukt ei while-løkke for å iterera over eit stadig mindre interval. I staden for løkka, kan ein bruka rekursjon; dvs. at funksjonen kallar seg sjølv."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def bisectR(f,l,u):\n",
    "    fl = f(l)\n",
    "    fu = f(u)\n",
    "    if fl*fu >= 0: raise Exception (\"Funksjonen byter ikkje forteikn på intervallet\")\n",
    "    m = (l+u)/2\n",
    "    fm = f(m)\n",
    "    if abs(l-u) < 0.01: return m\n",
    "    elif fm == 0: return m\n",
    "    elif fl*fm < 0: return bisectR(f,l,m)\n",
    "    elif fm*fu < 0: return bisectR(f,m,u)\n",
    "print ( bisectR(f,0,1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Legg merke til feilsjekken, som sikrar at me ikkje køyrer halveringsmetoden på eit intervall der funksjonen har same forteikn i båe endane.\n",
    "Det er ein god skikk å sikra seg mot alle tenkjelege feil på denne måten, men det er krevjande å tenkja på alt.  Dei fleste døma i dette kurset slurvar med denne typen feilsjekking.\n",
    "\n",
    "**Oppgåve:** Kva skjer om du droppar feilsjekken og køyrer t.d. ``bisectR(f,-1,0)``?\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def bisectRfeil(f,l,u):\n",
    "    fl = f(l)\n",
    "    fu = f(u)\n",
    "    m = (l+u)/2\n",
    "    fm = f(m)\n",
    "    if abs(l-u) < 0.01: return m\n",
    "    elif fm == 0: return m\n",
    "    elif fl*fm < 0: return bisectR(f,l,m)\n",
    "    elif fm*fu < 0: return bisectR(f,m,u)\n",
    "print ( bisectRfeil(f,-1,0))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Her fekk me ``None``; dvs. funksjonen gav ikkje nokon returverdi. Korleis kan det ha seg?\n",
    "Det må tyda at ingen av ``return``-satsane vart køyrde. \n",
    "Dersom du sjekkar utrekninga for hand, ser du at det gjev meining."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Utviklinga gjennom algoritmen\n",
    "\n",
    "Det er nyttig å sjå korleis estimatet utviklar seg steg for steg i algoritmen.\n",
    "I dømet under samlar me opp alle estimata i ei liste `r`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.0, -1.0, -0.5, -0.25, -0.125, -0.0625, -0.03125, -0.046875, -0.0546875, -0.05859375, -0.060546875, -0.0595703125]\n"
     ]
    }
   ],
   "source": [
    "def bisect2log(function,lower,upper):\n",
    "    r = []\n",
    "    while abs(lower-upper) > 0.001: \n",
    "        midpoint = (lower+upper)/2\n",
    "        fm = f(midpoint)\n",
    "        if f(lower)*fm < 0: upper = midpoint\n",
    "        elif fm*f(upper) < 0: lower = midpoint\n",
    "        r += [midpoint]\n",
    "    midpoint = (lower+upper)/2\n",
    "    return r\n",
    "xs = bisect2log(f,-2,2)\n",
    "print(xs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Merk at det siste talet i lista er det same estimatet som me har fått tidlegare.\n",
    "Skilnaden er at me no veit alle estimata på vegen.\n",
    "Dette kan me plotta:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7ff1513e6fd0>]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAD8CAYAAACfF6SlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvqOYd8AAAIABJREFUeJzt3Xl8XOV97/HPb2a0jmUtI+NdMwZkjEsgGOGwBNPUuIWQlCRNUnKb1tkKbbpQbtvUt9zk3tdtcsNtmlLStM11CI0hbdNAuYUGWjCGymbHxmGzseRNwvtosWRL1jKa5/4xIyHbWj2jObN836+XXprlaM5vvHzPo2d+5znmnENERAqLz+sCREQk8xT+IiIFSOEvIlKAFP4iIgVI4S8iUoAU/iIiBUjhLyJSgBT+IiIFSOEvIlKAAl4XMJ7a2loXiUS8LkNEJKds27atzTk3Z7Ltsjb8I5EIW7du9boMEZGcYmYtU9lO0z4iIgVI4S8iUoAU/iIiBUjhLyJSgBT+IiIFKC3hb2Y3mtkuM9ttZuvGeL7EzP45+fzLZhZJx34ncs/GppnehYhIzko5/M3MD/wNcBOwHPiMmS0/Y7MvAp3OuQuBe4D/k+p+J3PvpuaZ3oWISM5Kx8h/JbDbObfXOTcA/Bi45YxtbgE2JG8/DKw2M0vDvs/S1TvIvU8r+EVEJpKOk7wWAu+Oun8A+MB42zjnYmbWBYSAttEbmdltwG0AdXV10y7kno1Np434I+seB+CO1fXcuWbptF9PRCRfZdUZvs659cB6gIaGhmlfWf7ONUu5c81SVvzZRjp6Bth/981pr1FEJB+kY9rnILB41P1FycfG3MbMAkAl0J6GfY8pHCqfqZcWEckL6Qj/V4F6M1tiZsXArcBjZ2zzGLA2efuTwDPOuWmP7KcqEgoyq8Q/Uy8vcs7UhSbZIuXwd87FgN8FngR2Aj9xzr1tZv/LzH45udkPgJCZ7Qb+K3BWO2g6hUPl9AwM0Tc4NJO7EZk2daFJtkjLnL9z7gngiTMe+9qo233Ap9Kxr6mIhII4Bwc6e7nwvIpM7VZkXNET/bx1qAuA1vZe5laWUBLQb6finaz6wDdd6pJz/vvbFP6SebGhOO8cOcFrrZ281tLJM+8co7svNvL8qm89C0BZkY8LzpvF/MoyFlSWMq+yjAVVpcybXcqCqjLOm60DhMycvAz/SCgIwP72Ho8rkULQfrKf7a3Hea21k20tnbxxoItTySnH8ypKuOaCWlaEq7hsURW/uv4l/vyTl3Kkq4/DXac43NVHa3svL+9tP+0AMax2VslpB4R5laXMryxlfmUZ8ytLmTu7lOLA2LO392xsUouzjCsvw7+6vIiK0gAt7b1elyJ5Ziju2HXkBNtaO9ne0slrrZ3sT/47C/iM5Qtm86tXLubyuiquCFezsKqMM89n/HTD4rFempP9MY4kDwiHj/clvned4lBXH/vbe3hxTzsn+k8/QJglDxCVpckDQ+KgML+qjHs3NXPjJfOYVRKgojTArJIAAf/MLeelg01uycvwNzMioaBG/pKyzp4Btr/byWstiZH96+8ep2cgMaqvnVXCiroqbl1Zx4q6at63sJKy4omnae5YXT/uc7NKAlx4XsWEU5Un+gaTvzUkDwzH+zjS1cehrlPsjfbw/O52To46QNx075bTfr6syM+s0sTBoKIkwKzkQaGitGjkIJE4UBRNuJ3fd/YJ+vduas5o+GfyYJPpA1sm9peX4Q+Jjp83D3Z5XYZkudH/yYbijuZjJ0aC/rXWTvZGEwMIv8+4eH4Fv3LFIlbUVbOirprFNWeP6ieT6n/oitIiKkqLqJ879gHizLPch62qr+WKcA0n+wc52R+juy/Gyb4YJ/tjRE/0cLIvxon+xP2pNGGXF/tHfpuYVVpERUkiSn7nH1+jxO+jpMhHsd9HcSD55fe/dzvgoyT5dfo2o5/3J54/4/GAz0b+zDN5sMn0gS0T+8vb8I+Egvz7W0cYHIpTNIO/6kru6umPce+mZhywvbWT7a3HR0bNNcFiVtRV8SsrFnFFuJpLF1VSXpz9/12Gz3KHxPIm0z3LPR539A4OJQ8Mg6cdJE70DXIieftkX4wTfbGR34aGPf7GYSBxcCjy+xiIxRkYijMUT89pPWZQ7E8cOACu/MbT+M3w+xJfAd97t0d/BXyGz4yAP/n9tOd8+EZv47OR+8PbAHz9pztwQNw5nDv9e9yBG+s+791PbD/xz8Rn7vSns2T/v+ZzVBcqZyjuONh5ikht0OtyJIscPH6Kv39uHz9+NbEk1XefaWbZvNl87PIFI6P6cKh82qP6fODzWWI0XxIASqf1sxMdbGJDiYPAQCzx1R87/f7w7f7Y0HvPn7lNLM6W5iiv7O+kPxYHEi20ABfPq+CieRXE4o64c8SGkt/jjqHkVyzu6B+MT7jN6G1P9g/SNxgfeQ/3PbcPgGK/UVYcwAx8ZvgsMdVsnHF/1PO+5H077X7yZ3yJ+0e7+zja3X/anyfM3NpkeRv+ozt+FP4C8Pq7x/n+lr08/sZhRo+v4g52HO5mzfK5fGLFIs/qS7eJPl/ItIDfR8Dvo7w4tdf5vVHv6Vx+szlXmdxXpvaXx+Gf6PVXx09hG4o7Nu44yg+e28ur+zupKA1w26rzWXtNhAVVZRn/T51Jme68yaaDjUwub8N/TkUJZUV+dfwUqJ7+GA9vO8D9z++jpb2XRdVlfO0jy/n0lYuTUxqSbvl8sMn0gS0T+8vb/wVmRjhUTqtG/gXlcNcpNrzQwj++3EJ3X4wVdVWsu3EZa5bPHbPHXaPV3JXJg02mD2yZ2F/ehj8k5v2bj53wugzJgLcOdvGD5/bxb68fIu4cN10yny98cAlXhKsn/DmdlCSFKq/DPxwq55l3jjEUd2OelCK5LR53PLvrGN/fspeX9nYQLPbzG1dH+Py1ERbX6JoOIhPJ8/APMjAU53DXKRZVKwzyxamBIf7ltQPc/9w+9rb1sKCylD/98DJuXVnH7NIir8sTyQl5Hf6jO34U/rnv2Ik+HnyxhR+91EJn7yCXLqrkO5+5nJsumacT+USmKa/DP1z7Xq//tRfWelyNnKt3jnTzgy37ePRnhxiMx1lz8Vy+dN35XBmpLsgTsUTSIa/Df35yuVt1/OSO4bV2nHNsbm7jvi172dLcRlmRn1tXLuYL1y7RSXsiaZDX4e/zGXU15er1zyH3bmpmQVUp923ZR/Oxk8ydXcJXbryI/7KyjqpUTw8VkRF5Hf4A4ZpyneWbI556+wgAf/Ivb3Lx/Nn85acv4yOXLhj3YiUicu7yP/xDQZ7f04ZzTvPDWWqsZYh3Hu6mpb1XwS8yQ/I+/CO15fQNxjl2op+5s6e3SqHMPOfcyHK/q5edx6Z3juXtWjsi2STvh1Xh5OqemvrJPoNDcb7y8Bt899nd3HrlYv7vr1/hdUkiBSPvw3+4118f+maXnv4Yv/nAVh7adoA/uKGeb37ifQT8Pq21I5IheT/ts7CqjIDPaFH4Z422k/184Yev8tbBLr75iffxmZV1I89prR2RzMj78A/4fSysLmO/pn2yQkt7D79x/ysc7e5j/a83cMPyuV6XJFKQ8j78ITHvr5G/9944cJzP//2rxJ3jH3/zKlbUTbzipojMnLyf84fEvH9LWy8ugxdHltM9u+sYt65/ibJiPw//9jUKfhGPFUT4h0NBTvTH6Owd9LqUgvTQ1nf50oatLKkN8siXr+GCObO8Lkmk4BXEtM/ojp+aoJYIyBTnHH/z7G7+4qkmPnhhLX/32RVUaMllkaxQMCN/QPP+GTQUd3zt0bf5i6ea+PjlC7n/c1cq+EWySErhb2Y1ZrbRzJqT38+ayDWz95vZi2b2tpm9YWa/mso+z8XimjLMYH+bOn4yoW9wiN/+0TYefKmF37r+Ar79qcu0TINIlkn1f+Q6YJNzrh7YlLx/pl7gN5xzPwfcCPyVmVWluN9pKQn4WVBZppF/BhzvHeCz973Mxp1H+Z8fXc66m5bh0yU0RbJOqnP+twA/n7y9AfhP4E9Gb+Ccaxp1+5CZHQPmAMdT3Pe0hEPl6vWfYQePn2Lt/a/Q2t7Ldz+zgpsvne91SSIyjlRH/nOdc4eTt48AE56xY2YrgWJgT4r7nbZwKEhrh8J/puw83M0n/vZ5jnb38cAXVyr4RbLcpCN/M3samDfGU3eNvuOcc2Y2biO9mc0HHgTWOufi42xzG3AbQF1d3VibnLNIqJyOngG6Tg1SWaYPHtPphd1t3P7gNoIlAR7+rWu4aF6F1yWJyCQmDX/n3A3jPWdmR81svnPucDLcj42z3WzgceAu59xLE+xrPbAeoKGhIa1nZA13/LS29/K+RZXpfOmC9tjrh/jDn/yMJbVBfvj5lSyoKvO6JBGZglSnfR4D1iZvrwUePXMDMysG/h/wgHPu4RT3d84itVrdM93u27KX3/+n7VxeV81Dt1+j4BfJIamG/93AGjNrBm5I3sfMGszsvuQ2nwZWAZ8zs58lv96f4n6nra4mEf7q+EldPO74+k938PXHd3LTJfN44AsrqSzXVJpILkmp28c51w6sHuPxrcCXkrd/BPwolf2kQ3lxgPMqStTxk6L+2BB//NAbPPb6IdZeHeZrH/05/GrlFMk5BbG8w7BIKEirwv+cdfcN8lsPbuOFPe2su2kZt686X9dFFslRBRX+4VA5jU1Rr8vISUe7+1h7/yvsPnaSv/z0ZXxixSKvSxKRFBRU+Edqgzy07QC9AzHKiwvqrafkq//6Js+8E+V47wD3f+5KVi2d43VJIpKiglpwJRwa/tBXUz9Tta2lgwdfaqU/Fuefb79awS+SJwor/Gu0uud0vNvRy2fvewWAR377Gi5ZqPMjRPJFQc191I2s66+R/2Tu2djEvZuaR+6v+tazANyxul4XWRfJAwUV/pVlRdQEizXtMwXXXzSHezc18/ur6/nOpmb2332z1yWJSBoV1LQPJOb9Ne0zMecSJ3HNqSjh9lXne12OiMyAggv/SCiokf8knnjzCK+1HuePfnEpwZIAd6yu97okEUmzggv/cKicQ12n6Bsc8rqUrNQfG+Lu/9jJsnkVfPKKxQCa4xfJQwUZ/s7BgU6N/sey4YX9vNtxiv9+83It2yCSxwow/BPtnrqe79k6egb462d286GL5vDB+lqvyxGRGVRw4R9Jhn+Lrup1lnufbqJ3YIg//fDFXpciIjOs4MK/uryIitKAOn7OsCd6kn94uZXPrFxM/VxdiUsk3xVc+JsZkVBQJ3qd4ZtPvENpkZ8/uEEf7ooUgoILf1Cv/5le2NPG0zuP8uUPXUDtrBKvyxGRDCjY8D/QeYrBoTGvI19Q4nHHNx7fycKqMr5w7RKvyxGRDCnQ8A8yFHcc7DzldSmee2T7Qd4+1M1XbryI0iK/1+WISIYUZPir4yehdyDGt558h8sWV/HLly3wuhwRyaACDX9dzB3g+5v3cbS7n6/efLEuxyhSYAoy/OdUlFBW5C/oE72OdvfxvcY9fPh982iI1HhdjohkWEGGv5kVfMfPt5/aRSwe509uXOZ1KSLigYIMf0h0/Owv0PDfcaibh7Yd4HPXREaWuxCRwlKw4R8JBXm34xRDced1KRnlnOMbT+ygsqyI3/2QlmoWKVQFG/7hUJCBoThHuvu8LiWjnt11jOd3t3PH6noqy4u8LkdEPFKw4T/S8dNWOFM/g0NxvvH4Ts6vDfLZq8JelyMiHirY8A/XJpd2LqA1fn78Sit7oj2su2kZRf6C/asXEQo4/OfPLqU44CuYjp/uvkHuebqZDyypYc3yuV6XIyIeK9jw9/mMxdVlBdPx87fP7qGzd4CvfmS5TugSkcINfyici7m/29HL/c/t4+OXL+SShZVelyMiWaCgwz+cDH/n8rvd88+f3IXPB3/8Sxd5XYqIZImUwt/Masxso5k1J79XT7DtbDM7YGbfTWWf6RSpLefU4BDRE/1elzJjXmvt5N9eP8Rt153P/Moyr8sRkSyR6sh/HbDJOVcPbEreH8+fAZtT3F9ajVzMPU+nfpxzfP2nO5hTUcLt11/gdTkikkVSDf9bgA3J2xuAj421kZldAcwFnkpxf2k13Oufrx/6PvHmEV5rPc4frllKsCTgdTkikkVSDf+5zrnDydtHSAT8aczMB3wb+KPJXszMbjOzrWa2NRqNplja5BZUleH3WV62e/bHhrj7P3aybF4Fn2pY7HU5IpJlJh0OmtnTwLwxnrpr9B3nnDOzsT45/TLwhHPuwGQths659cB6gIaGhhn/FLbI72NRdVleTvtseGE/73ac4sEvrsTvU2uniJxu0vB3zt0w3nNmdtTM5jvnDpvZfODYGJtdDVxnZl8GZgHFZnbSOTfR5wMZEw4Fac2z8O/oGeCvn9nNz180h+vq53hdjohkoVSnfR4D1iZvrwUePXMD59yvOefqnHMRElM/D2RL8ENi3n9/e09etXt+Z1MzvQND3PXhi70uRUSyVKrhfzewxsyagRuS9zGzBjO7L9XiMiEcCnKiL0Zn76DXpaTFnuhJfvRSC7deuZj6uRVelyMiWSqlFhDnXDuweozHtwJfGuPxHwI/TGWf6Ta646cmWOxxNan75hPvUFrk5841S70uRUSyWEGf4QuJK3pBflzM/YU9bTy98yhf/tAF1M4q8bocEcliBR/+i6rLMSPnL+Yejzu+8fhOFlaV8YVrl3hdjohkuYIP/9IiPwsqy2jtyO3wf2T7Qd4+1M1XbryI0iK/1+WISJYr+PCH3L+Ye+9AjG89+Q6XLa7io5cu8LocEckBCn/eW90zV31/8z6Odvfz1ZsvxqcTukRkChT+JDp+OnoG6DqVe+2eR7v7+F7jHm66ZB4NkRqvyxGRHKHw572On1w80/fbT+0iFo+z7qZlXpciIjlE4c/opZ1za95/x6FuHtp2gLVXR0beg4jIVCj8GTXyz6GOH+cc33hiB5VlRfzeL9R7XY6I5BiFP1BeHOC8ihL2t+XOyP/ZXcd4fnc7d6yup7K8yOtyRCTHKPyTculi7rGhOP/7iXdYUhvk1z4Q9rocEclBCv+kXOr1/80HtrL72EnW3bSM4oD+CkVk+pQcSeFQOcdO9NM7EPO6lAl19w3y7K4oH1hSwy8uP+vCaSIiU6LwTxrulsn2D32/9597APjvNy9nsiujiYiMR1f1TooMt3u29bJs3myPqznbPRubuHdT88j9j373OQDuWF2v5ZtFZNoU/kl1Wb60851rlvLh983nl/5qMwD7777Z44pEJJdp2iepsqyImmBxVl/MvbFprEski4hMn0b+o4RD5Vk78gfY3NTG0rmzuOmS+V6XIiI5TiP/UcI15Vnb6987EOOVfR1cv3SO5vhFJGUK/1HCoSCHuk7RHxvyupSzvLS3nYGhOKuWzvG6FBHJAwr/USK15TgH73ac8rqUs2xuaqO0yMeVWrZZRNJA4T/KcK9/Ns77NzZFufr8kC7RKCJpofAfZaTXP8vm/Vvbe9nX1qMpHxFJG4X/KNXlRVSUBrJu5N/YHAXgeoW/iKSJwn8UM0su8JZdI//GXVEW15SxpFYXbBGR9FD4nyEcCtKaRSP/gVicF/e0sap+jtbyEZG0UfifIRIq50DnKQaH4l6XAsC2lk56BoY05SMiaaXwP0M4FCQWdxw6nh3tno1NUQI+45oLa70uRUTyiML/DNnW8dPYFOWKcDWzSrQSh4ikj8L/DJEsWt3zWHcfOw93c/1FmvIRkfRKKfzNrMbMNppZc/J79Tjb1ZnZU2a208x2mFkklf3OpDkVJZQV+dnf5v3If3NzG6AWTxFJv1RH/uuATc65emBT8v5YHgC+5Zy7GFgJZO3axMPtnq0d3o/8G5ui1M4q4eIsvLiMiOS2VMP/FmBD8vYG4GNnbmBmy4GAc24jgHPupHPO+2H1BLKh138o7niuOcqqpbX4fGrxFJH0SjX85zrnDidvHwHGuqL4UuC4mT1iZtvN7FtmltUL1ERCQVrbexmKO89qePNgF529g5ryEZEZMWkLiZk9Dcwb46m7Rt9xzjkzGystA8B1wOVAK/DPwOeAH4yxr9uA2wDq6uomK23GhENBBobiHOnuY2FVmSc1NO6KYgbX1Sv8RST9Jg1/59wN4z1nZkfNbL5z7rCZzWfsufwDwM+cc3uTP/OvwFWMEf7OufXAeoCGhgbPht0jHT9tPZ6F/+bmKJcurKQmWOzJ/kUkv6U67fMYsDZ5ey3w6BjbvApUmdnwEPYXgB0p7ndGDV/M3at5/67eQba3dmrKR0RmTKrhfzewxsyagRuS9zGzBjO7D8A5NwT8EbDJzN4EDPh+ivudUfMryyj2+2jxqOPnud1txB1awllEZkxKp40659qB1WM8vhX40qj7G4FLU9lXJvl9xuKaMlo86vXf3BSlojTA+xdXebJ/Ecl/OsN3HJFQkP0enOXrnKOxKcp19bUE/PrrEZGZoXQZRzgUpKW9F+cy+7lz09GTHOnuY5W6fERkBin8xxGpLefU4BDRE/0Z3e/mpsRVuzTfLyIzSeE/jroabzp+GpuiLJ07iwUetZiKSGFQ+I9jeGnnTK7u2TsQ45V9HZryEZEZp/Afx8LqMvw+oyWDI/+X93YwMBTXEs4iMuMU/uMo8vtYVF2W0Y6fxqYopUU+rozUZGyfIlKYFP4TGO74yZTGpihXnR+itCir170TkTyg8J9AJFTO/vaejLR7trb3sq+tR0s6iEhGKPwnUFdTzom+GJ29gzO+r8bmRIunwl9EMkHhP4FMdvw07oqyuKaMJbXBGd+XiIjCfwKR2uGLuc/svP9ALM6Le9pYVT8HM121S0RmnsJ/AouqyzFjxjt+trV00jMwpCkfEckYhf8ESov8LKgsm/GRf2NTlIDPuObC2hndj4jIMIX/JMLJjp+ZtLkpyhXhamaVpLTCtojIlCn8JxEOlc/oyP9Ydx87DnfrrF4RySiF/yTCoSAdPQN0981Mu+fm5jZALZ4iklkK/0kMX8y9dYZG/5ubotTOKuHiebNn5PVFRMai8J9EONnrPxPz/kNxx5bmKKuW1uLzqcVTRDJH4T+JcGjmev3fPNhFZ++gpnxEJOMU/pMoLw5wXkUJ+9vSP/Lf3BTFDK7T+v0ikmEK/ymYqY6fxqYoly6spCZYnPbXFhGZiMJ/CsKhIC0d6R35d/UOsr21U1M+IuIJhf8URELlHO3up3cglrbXfH5PG3GnC7WLiDcU/lMw3PHT2pG+qZ/GXVEqSgO8f3FV2l5TRGSqFP5TMLy08/629IS/c47GpijX1dcS8OuvQEQyT8kzBXUj7Z7pmfdvPnaSI919rFKXj4h4ROE/BZVlRVSXF7E/TR0/jbsSV+3SfL+IeEXhP0XhUJDWNHX8NDZFWTp3FguqytLyeiIi06Xwn6JIqDwtc/69AzFe2dehKR8R8ZTCf4rCoSCHuk7RHxtK6XVe3tvBwFBcSziLiKdSCn8zqzGzjWbWnPxePc52f25mb5vZTjP7juXghWojteU4B+92nErpdRqbopQW+bgyUpOmykREpi/Vkf86YJNzrh7YlLx/GjO7BrgWuBS4BLgSuD7F/WbccK9/qh0/m5uiXHV+iNIifzrKEhE5J6mG/y3AhuTtDcDHxtjGAaVAMVACFAFHU9xvxoVrEu2eqXT8tLb3sretR0s6iIjnUg3/uc65w8nbR4C5Z27gnHsReBY4nPx60jm3c6wXM7PbzGyrmW2NRqMplpZeNcFiKkoCtKYw8m9sTrwnhb+IeG3SK4ab2dPAvDGeumv0HeecMzM3xs9fCFwMLEo+tNHMrnPObTlzW+fcemA9QENDw1mv5SUzI1xbntLIv3FXlEXVZSypDaaxMhGR6Zs0/J1zN4z3nJkdNbP5zrnDZjYfODbGZh8HXnLOnUz+zL8DVwNnhX+2C4eCvH2w65x+diAW58U9bXzs8oXk4OfdIpJnUp32eQxYm7y9Fnh0jG1agevNLGBmRSQ+7B1z2ifbRULlHOg8xeBQfNo/u62lk56BIU35iEhWSDX87wbWmFkzcEPyPmbWYGb3Jbd5GNgDvAm8DrzunPu3FPfriXAoSCzuOHR8+u2ejU1RAj7j6gtCM1CZiMj0TDrtMxHnXDuweozHtwJfSt4eAm5PZT/ZYnTHz3Dr51RtbopyRbiaitKimShNRGRadIbvNESSH9ROt+PnWHcfOw5366xeEckaCv9pOK+ihNIi37Q7fjY3twFoPR8RyRoK/2kwMyKh4LTP8t3cFKV2VgnL58+eocpERKZH4T9N4dD0ev2H4o4tzVFWLa3F51OLp4hkB4X/NEVCQVo7eonHp3YO2psHu+jsHVSLp4hkFYX/NNWFyhmIxTnS3Tel7Tc3RTGDD15YO8OViYhMncJ/mkYu5j7Fef/GpiiXLqwkNKtkJssSEZkWhf80hUcu5j75vH9X7yDbWzs15SMiWUfhP03zK8so9vumNPJ/fk8bcacLtYtI9lH4T5PfZyyuKaNlCtfzbdwVpaI0wPsXV2WgMhGRqVP4n4NIKDjpyN85R2NTlOvqawn49ccsItlFqXQO6kLltHb04tz47Z7Nx05ypLtPZ/WKSFZS+J+DSChI78AQ0ZP9427TuCtx1S7N94tINlL4n4OpdPw0NkVZOncWC6rKMlWWiMiUKfzPwUivf9vY8/69AzFe2dehKR8RyVoK/3OwsLoMv8/GHfm/vLeDgaG4lnAWkayl8D8HRX4fi6rLxu34aWyKUlrk48pITYYrExGZGoX/OaqrSXT8jGVzU5Srzg9RWuTPcFUiIlOj8D9HkVCQfW09Z7V7trb3sretR0s6iEhWU/ifo3ConBN9MY73Dp72eGNzosVT4S8i2Uzhf47GW91zc1OURdVlLKmd3gXeRUQySeF/jiK1Z/f6D8TivLC7jeuXzsFMV+0Skeyl8D9Hi6rLMTt95L+tpZOegSFN+YhI1lP4n6PSIj/zZ5fSOmrkv7k5SsBnXH1ByMPKREQmp/BPQfiM1T0bd0W5IlxNRWmRh1WJiExO4Z+CSG35yJz/sRN97DjcrbN6RSQnKPxTEA4Fae8ZoLtvkC1NbQBaz0dEcoLCPwWR5Oqere29NDZFqZ1VwvL5sz0C5q77AAAE6klEQVSuSkRkcgr/FISTvf5723rY0hxl1dJafD61eIpI9lP4p6CuJjHyf/yNQ3T2DqrFU0RyRkrhb2afMrO3zSxuZg0TbHejme0ys91mti6VfWaTYEmAORUlbNxxFDP44IW1XpckIjIlqY783wI+AWwebwMz8wN/A9wELAc+Y2bLU9xv1oiEyok7uHRhJaFZJV6XIyIyJSmFv3Nup3Nu1ySbrQR2O+f2OucGgB8Dt6Sy32wyPO+va/WKSC7JxJz/QuDdUfcPJB/LC8MdP5rvF5FcEphsAzN7Gpg3xlN3OeceTWcxZnYbcBtAXV1dOl867e7Z2MS9m5pH7n/yey8CcMfqeu5cs9SrskREpmTS8HfO3ZDiPg4Ci0fdX5R8bKx9rQfWAzQ0NLixtskWd65ZOhLykXWPs//umz2uSERk6jIx7fMqUG9mS8ysGLgVeCwD+xURkXGk2ur5cTM7AFwNPG5mTyYfX2BmTwA452LA7wJPAjuBnzjn3k6t7Oxyx+p6r0sQEZkWO/MatNmioaHBbd261esyRERyipltc86Ne97VMJ3hKyJSgBT+IiIFSOEvIlKAFP4iIgVI4S8iUoCyttvHzKJASwovUQu0pamcbKP3lrvy+f3pvWWHsHNu0vVmsjb8U2VmW6fS7pSL9N5yVz6/P7233KJpHxGRAqTwFxEpQPkc/uu9LmAG6b3lrnx+f3pvOSRv5/xFRGR8+TzyFxGRceRd+OfrxeIBzGyxmT1rZjvM7G0zu8PrmtLNzPxmtt3Mfup1LelkZlVm9rCZvWNmO83saq9rSiczuzP5b/ItM/snMyv1uqZzZWb3m9kxM3tr1GM1ZrbRzJqT36u9rDEd8ir88/1i8UAM+EPn3HLgKuB38uz9AdxBYunvfHMv8B/OuWXAZeTRezSzhcDvAw3OuUsAP4nrduSqHwI3nvHYOmCTc64e2JS8n9PyKvzJ84vFO+cOO+deS94+QSJA8uZ6yGa2CLgZuM/rWtLJzCqBVcAPAJxzA865495WlXYBoMzMAkA5cMjjes6Zc24z0HHGw7cAG5K3NwAfy2hRMyDfwj+vLxY/mplFgMuBl72tJK3+CvgKEPe6kDRbAkSBv09Oad1nZkGvi0oX59xB4C+AVuAw0OWce8rbqtJurnPucPL2EWCul8WkQ76Ff0Ews1nAvwB/4Jzr9rqedDCzjwDHnHPbvK5lBgSAFcDfOecuB3rIg2mDYcn571tIHOQWAEEz+6y3Vc0cl2iRzPk2yXwL/ylfLD5XmVkRieD/B+fcI17Xk0bXAr9sZvtJTNf9gpn9yNuS0uYAcMA5N/xb2sMkDgb54gZgn3Mu6pwbBB4BrvG4pnQ7ambzAZLfj3lcT8ryLfzz+mLxZmYk5o13Ouf+0ut60sk599+cc4uccxESf2/POOfyYvTonDsCvGtmFyUfWg3s8LCkdGsFrjKz8uS/0dXk0QfaSY8Ba5O31wKPelhLWgS8LiCdnHMxMxu+WLwfuD/PLhZ/LfDrwJtm9rPkY3/qnHvCw5pkan4P+IfkoGQv8HmP60kb59zLZvYw8BqJjrTt5PAZsWb2T8DPA7VmdgD4H8DdwE/M7IskVhv+tHcVpofO8BURKUD5Nu0jIiJToPAXESlACn8RkQKk8BcRKUAKfxGRAqTwFxEpQAp/EZECpPAXESlA/x/VNn4o0C+zdQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "n = len(xs)\n",
    "x = range(n)\n",
    "plt.plot(x,xs,\"+-\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Oppgåve\n",
    "\n",
    "1. Korleis kan du auka presisjonen i halveringsmetoden, og få fleire riktige desimalar i svaret?\n",
    "1. Gå vidare til dokumentet om Sekantmetoden.  Det er ein raskara algoritme i dei fleste tilfelle. Dokumentet gjev hovudstega, men du må fullføra algoritmen vha. idéane som me har presentert her."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Oppsummering\n",
    "\n",
    "Det viktigaste i dette dømet er \n",
    "1. Vilkårssatsen\n",
    "1. Løkker\n",
    "1. Algoritmen for halveringsmetoden.\n",
    "Me har òg vist rekursjon.  Alt som me kan gjera med rekursjon, kan me òg gjera med løkker og vise versa.\n",
    "Det er mykje smak og behag som avgjer kva me vel."
   ]
  }
 ],
 "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": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}
