examples/tough_heat/joint_inversion_example_tough.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# pyPCGA-TOUGH tutorial example - permeability estimation using pressure and temperature measurements\n",
    "\n",
    "## Problem description\n",
    "\n",
    "+ Domiain : 300 m x 100 m x 100 m \n",
    "+ 3000 (nx, ny, mz = 30, 30, 10) log-normal permability using Gaussian covariance kernel with scale parameter of [100,100,10] (m) \n",
    "+ Boundary conditions: injection-extraction system, warm water injected\n",
    "+ Simulation time: 374 days\n",
    "+ Number of unknowns: 3,000 permeabilities (pmx values)\n",
    "+ Measurements: pressure and temperature collected every 5 days at 128 monitoring locations between the injection and extraction well"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Test environment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.7.0 (default, Jun 28 2018, 13:15:42) \n",
      "[GCC 7.2.0]\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "print(sys.version)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linux-4.15.0-1021-aws-x86_64-with-debian-buster-sid\n"
     ]
    }
   ],
   "source": [
    "import platform\n",
    "print(platform.platform()) # ubuntu 18.04"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.15.1\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "print(np.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# of physcial CPU cores 36\n"
     ]
    }
   ],
   "source": [
    "import psutil\n",
    "print('# of physcial CPU cores %s' % (psutil.cpu_count(logical=False)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import tough\n",
    "from pyPCGA import PCGA\n",
    "import math\n",
    "import warnings; warnings.simplefilter('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- forward model information: model domain, discretization, observation location and interval\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# forward model information: model domain, discretization, observation location and interval\n",
    "nx = [30, 10, 10]\n",
    "dx = [10., 10., 2.]\n",
    "\n",
    "# monitoring indices \n",
    "xlocs = np.arange(5,21,2)\n",
    "ylocs = np.array([1,3,5,7])\n",
    "zlocs = np.array([3,5,7,9])\n",
    "\n",
    "forward_model_params = {'nx':nx,'dx':dx, 'deletedir':False, 'xlocs': xlocs, 'ylocs':ylocs, 'zlocs':zlocs, \\\n",
    "    'obs_type':['Gas Pressure','Temperature'],'t_obs_interval':86400.*5.,\\\n",
    "    'max_timesteps': 9000, 'tstop': 0.32342126E+08, 'const_timestep': 6, 'max_timestep':86400, \\\n",
    "    'absolute_error': 1, 'relative_error': 5.e-6, 'print_interval': 9000, 'timestep_reduction': 3., \\\n",
    "    'gravity': 9.81,'default_incons': [100.e4, 10], \\\n",
    "    'multi_params':{'num_components': 1, 'num_equations':2, 'num_phases':2, 'num_secondary_parameters':6},\n",
    "    'solver_params':{'type': 5, 'z_precond':1,'o_precond': 0, 'relative_max_iterations':.8,'closure':1.e-7 },\n",
    "    'output_times_params': {'num_times_specified':2, 'time': [0.8640E+04, 0.32342126E+08]}}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- true pmx values for comparison purpose"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "s_true = np.loadtxt(\"true_30_10_10_gau.txt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "s_true3d = s_true.reshape([nx[2],nx[1],nx[0]])\n",
    "for i in range(1,10,3):\n",
    "    plt.figure()\n",
    "    plt.pcolormesh(s_true3d[i,:,:],vmin=-2.0,vmax=1.5, cmap=plt.get_cmap('jet'))\n",
    "    plt.title('true ln(pmx) in layer %0d' %(i))\n",
    "    ax = plt.gca()\n",
    "    ax.set_aspect('equal', 'box')\n",
    "    ax.xaxis.set_ticks([])\n",
    "    ax.yaxis.set_ticks([])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- PCGA parameters (m: # of unknowns, N, xmin, xmax: structure grid parameters, prior information)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dimension of unknown\n",
    "m = nx[0]*nx[1]*nx[2]\n",
    "N = np.array([nx[0],nx[1],nx[2]])\n",
    "xmin = np.array([0.,0.,0.])\n",
    "xmax = np.array([300.,100.,20.])\n",
    "pts = None # for regular grids, you don't need to specify pts. \n",
    "\n",
    "# covairance kernel and scale parameters\n",
    "prior_std = 0.7\n",
    "prior_cov_scale = np.array([100.0,100.0,10.0])\n",
    "\n",
    "def kernel(r): return (prior_std ** 2) * np.exp(-r**2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "+ load noisy pressure and temperature measurements\n",
    "\n",
    "measurements were generated as\n",
    "\n",
    "```python\n",
    "\n",
    "obs_pres = simul_obs_pres + 10000.*np.random.randn(n_pres,1)  # error std = 1000 Pa for pressure\n",
    "obs_temp = simul_obs_temp + .5*np.random.randn(n_temp,1)  # error std = 0.5 C for pressure\n",
    "```\n",
    "\n",
    "where simul_obs_pres and simul_obs_temp are simulated pressure and temperature at the monitoring ports, respectively"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load noisy observations\n",
    "obs = np.loadtxt('obs.txt')\n",
    "nobs = obs.shape[0]\n",
    "\n",
    "# joint inversion prior error level\n",
    "std_obs = np.ones_like(obs)\n",
    "std_obs[:7400]=1000.\n",
    "std_obs[7400:]=0.5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "+ Define a wrapper for a black-box forward model input in pyPCGA\n",
    "\n",
    "#### 1. Note that one should follow this wrapper format to work with pyPCGA\n",
    "#### 2. If parallelization == True, it should take multiple columns of s and run the forward problem independently in parallel\n",
    "#### 3. Please see tough.py as a template for inplementation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# forward model wrapper for pyPCGA\n",
    "def forward_model(s, parallelization, ncores=None):\n",
    "    model = tough.Model(forward_model_params)\n",
    "\n",
    "    if parallelization:\n",
    "        simul_obs = model.run(s, parallelization, ncores)\n",
    "    else:\n",
    "        simul_obs = model.run(s, parallelization)\n",
    "    return simul_obs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- PCGA inversion parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "params = {'R': std_obs** 2, 'n_pc': 30,\n",
    "        'maxiter': 3, 'restol': 0.6,\n",
    "        'matvec': 'FFT', 'xmin': xmin, 'xmax': xmax, 'N': N,\n",
    "        'prior_std': prior_std, 'prior_cov_scale': prior_cov_scale,\n",
    "        'kernel': kernel, 'post_cov': \"diag\",\n",
    "        'precond': True, 'LM': True,\n",
    "        'parallel': True, 'linesearch': True, 'precision': 2e-3,\n",
    "        'forward_model_verbose': False, 'verbose': False,\n",
    "        'iter_save': True, 'ncores':6}\n",
    "\n",
    "# params['objeval'] = False, if true, it will compute accurate objective function"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- initial guess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "s_init = np.mean(s_true) * np.ones((m, 1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inversion \n",
    "\n",
    "- initialize pyPCGA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "##### PCGA Inversion #####\n",
      "##### 1. Initialize forward and inversion parameters\n",
      "------------ Inversion Parameters -------------------------\n",
      "   Number of unknowns                               : 3000\n",
      "   Number of observations                           : 14800\n",
      "   Number of principal components (n_pc)            : 30\n",
      "   Prior model                                      : def kernel(r): return (prior_std ** 2) * np.exp(-r**2)\n",
      "\n",
      "   Prior variance                                   : 4.900000e-01\n",
      "   Prior scale (correlation) parameter              : [100. 100.  10.]\n",
      "   Posterior cov computation                        : diag\n",
      "   Posterior variance computation                   : Approx.\n",
      "   Number of CPU cores (n_core)                     : 6\n",
      "   Maximum GN iterations                            : 3\n",
      "   machine precision (delta = sqrt(precision))      : 2.000000e-03\n",
      "   Tol for iterations (norm(sol_diff)/norm(sol))    : 6.000000e-01\n",
      "   Levenberg-Marquardt (LM)                         : True\n",
      "   LM solution range constraints (LM_smin, LM_smax) : None, None\n",
      "   Line search                                      : True\n",
      "-----------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# initialize\n",
    "prob = PCGA(forward_model, s_init, pts, params, s_true, obs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- run pyPCGA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "##### 2. Construct Prior Covariance Matrix\n",
      "- time for covariance matrix construction (m = 3000) is 0 sec\n",
      "##### 3. Eigendecomposition of Prior Covariance\n",
      "- time for eigendecomposition with k = 30 is 0 sec\n",
      "- 1st eigv : 369.519, 30-th eigv : 2.52487, ratio: 0.00683285\n",
      "##### 4. Start PCGA Inversion #####\n",
      "-- evaluate initial solution\n",
      "obs. RMSE (norm(obs. diff.)/sqrt(nobs)): 154474, normalized obs. RMSE (norm(obs. diff./sqrtR)/sqrt(nobs)): 154.602\n",
      "***** Iteration 1 ******\n",
      "computed Jacobian-Matrix products in 139.933877 secs\n",
      "solve saddle point (co-kriging) systems with Levenberg-Marquardt\n",
      "evaluate LM solutions\n",
      "LM solution evaluted\n",
      "- Geostat. inversion at iteration 1 is 164 sec\n",
      "== iteration 1 summary ==\n",
      "= objective function is 6.382210e+07, relative L2-norm diff btw sol 0 and sol 1 is 24.3275\n",
      "= L2-norm error (w.r.t truth) is 79.736, obs. RMSE is 92833.5, obs. normalized RMSE is 92.8688\n",
      "- save results in text at iteration 1\n",
      "***** Iteration 2 ******\n",
      "computed Jacobian-Matrix products in 137.223684 secs\n",
      "solve saddle point (co-kriging) systems with Levenberg-Marquardt\n",
      "evaluate LM solutions\n",
      "LM solution evaluted\n",
      "- Geostat. inversion at iteration 2 is 164 sec\n",
      "== iteration 2 summary ==\n",
      "= objective function is 6.278530e+05, relative L2-norm diff btw sol 1 and sol 2 is 0.794698\n",
      "= L2-norm error (w.r.t truth) is 84.7643, obs. RMSE is 9132.37, obs. normalized RMSE is 9.21096\n",
      "- save results in text at iteration 2\n",
      "***** Iteration 3 ******\n",
      "computed Jacobian-Matrix products in 137.983859 secs\n",
      "solve saddle point (co-kriging) systems with Levenberg-Marquardt\n",
      "evaluate LM solutions\n",
      "LM solution evaluted\n",
      "- Geostat. inversion at iteration 3 is 162 sec\n",
      "== iteration 3 summary ==\n",
      "= objective function is 4.080623e+05, relative L2-norm diff btw sol 2 and sol 3 is 0.503219\n",
      "= L2-norm error (w.r.t truth) is 92.882, obs. RMSE is 7349.22, obs. normalized RMSE is 7.42545\n",
      "- save results in text at iteration 3\n",
      "start posterior variance computation\n",
      "posterior diag. computed in 3.246672 secs\n",
      "------------ Inversion Summary ---------------------------\n",
      "** Found solution at iteration 3\n",
      "** Solution obs. RMSE 7349.22 , initial obs. RMSE 154474, where RMSE = (norm(obs. diff.)/sqrt(nobs)), Solution obs. nRMSE 7.42545, init. obs. nRMSE 154.602\n",
      "** Final objective function value is 4.080623e+05\n",
      "** Final predictive model checking Q2, cR is 8.399792e-01, 8.197448e+02\n",
      "** Total elapsed time is 515.687284 secs\n",
      "----------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# run inversion\n",
    "s_hat, simul_obs, post_diagv, iter_best = prob.Run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "post_std = np.sqrt(post_diagv)\n",
    "\n",
    "s_hat3d = s_hat.reshape([nx[2],nx[1],nx[0]])\n",
    "post_std = post_std.reshape([nx[2],nx[1],nx[0]])\n",
    "\n",
    "from mpl_toolkits.axes_grid1 import make_axes_locatable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(1,10,3):\n",
    "    fig, ax = plt.subplots(nrows=2,ncols=1,sharex=True)\n",
    "    ax[0].pcolormesh(s_true3d[i,:,:],vmin=-2.0,vmax=1.5, cmap=plt.get_cmap('jet'))\n",
    "    ax[0].set_title('true ln(pmx) in layer %0d' %(i))\n",
    "    ax[0].xaxis.set_ticks([])\n",
    "    ax[0].yaxis.set_ticks([])\n",
    "    ax[1].pcolormesh(s_hat3d[i,:,:],vmin=-2.0,vmax= 1.5, cmap=plt.get_cmap('jet'))\n",
    "    ax[1].set_title('estimated ln(pmx) in layer %0d' %(i))\n",
    "    ax[1].xaxis.set_ticks([])\n",
    "    ax[1].yaxis.set_ticks([])\n",
    "    \n",
    "    fig = plt.figure()\n",
    "    ax = plt.gca()\n",
    "    im = plt.pcolor(post_std[i,:,:], cmap=plt.get_cmap('jet'))\n",
    "    plt.title('Uncertainty (posterior std) in lnK estimate, layer %d' % (i))\n",
    "    divider = make_axes_locatable(ax)\n",
    "    cax = divider.append_axes(\"right\", size=\"5%\", pad=0.05)\n",
    "    plt.colorbar(im, cax=cax)\n",
    "    ax.set_aspect('equal', 'box')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[None, None]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "interval_pres = 1000000\n",
    "interval_temp = 10\n",
    "nobs = obs.shape[0]\n",
    "fig, ax = plt.subplots(nrows=1,ncols=2)\n",
    "\n",
    "obs_pres = obs[:7400]\n",
    "simul_obs_pres = simul_obs[:7400]\n",
    "obs_temp = obs[7400:]\n",
    "simul_obs_temp = simul_obs[7400:]\n",
    "\n",
    "from matplotlib.ticker import FormatStrFormatter\n",
    "\n",
    "ax[0].plot(obs_pres,simul_obs_pres,'.')\n",
    "ax[0].set_xlabel('observation')\n",
    "ax[0].set_ylabel('simulation')\n",
    "ax[0].set_title('pressure')\n",
    "minobs = np.vstack((obs_pres.reshape(-1), simul_obs_pres.reshape(-1))).reshape(-1).min()\n",
    "maxobs = np.vstack((obs_pres.reshape(-1), simul_obs_pres.reshape(-1))).reshape(-1).max()\n",
    "xmin, xmax = math.floor(minobs/interval_pres)*interval_pres, math.ceil(maxobs/interval_pres)*interval_pres\n",
    "ax[0].plot(np.linspace(xmin, xmax, 20), np.linspace(xmin, xmax, 20), 'k-')\n",
    "ax[0].set_xlim([xmin,xmax])\n",
    "ax[0].set_ylim([xmin,xmax])\n",
    "ax[0].xaxis.set_ticks(np.linspace(xmin,xmax,int(xmax/interval_pres)+1))\n",
    "ax[0].yaxis.set_ticks(np.linspace(xmin,xmax,int(xmax/interval_pres)+1))\n",
    "ax[0].xaxis.set_major_formatter(FormatStrFormatter('%.0e'))\n",
    "ax[0].yaxis.set_major_formatter(FormatStrFormatter('%.0e'))\n",
    "ax[0].set(adjustable='box', aspect='equal')\n",
    "\n",
    "ax[1].plot(obs_temp,simul_obs_temp,'.')\n",
    "ax[1].set_xlabel('observation')\n",
    "ax[1].set_ylabel('simulation')\n",
    "ax[1].set_title('temperature')\n",
    "minobs = np.vstack((obs_temp.reshape(-1), simul_obs_temp.reshape(-1))).reshape(-1).min()\n",
    "maxobs = np.vstack((obs_temp.reshape(-1), simul_obs_temp.reshape(-1))).reshape(-1).max()\n",
    "xmin, xmax = math.floor(minobs/interval_temp)*interval_temp, math.ceil(maxobs/interval_temp)*interval_temp\n",
    "ax[1].plot(np.linspace(xmin, xmax, 20), np.linspace(xmin, xmax, 20), 'k-')\n",
    "ax[1].set_xlim([xmin,xmax])\n",
    "ax[1].set_ylim([xmin,xmax])\n",
    "ax[1].xaxis.set_ticks(np.linspace(xmin,xmax,int(xmax/interval_temp)+1))\n",
    "ax[1].yaxis.set_ticks(np.linspace(xmin,xmax,int(xmax/interval_temp)+1))\n",
    "ax[1].set(adjustable='box', aspect='equal')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.9, 3.1, 316967.16050247126, 82164329.01137531)"
      ]
     },
     "execution_count": 20,
     "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": [
    "plt.semilogy(np.linspace(1,len(prob.objvals),len(prob.objvals)), prob.objvals, 'r-')\n",
    "plt.xticks(np.linspace(1,len(prob.objvals),len(prob.objvals)))\n",
    "plt.title('obj values over iterations')\n",
    "plt.axis('tight')"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}