{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "
\n", "\n", "# Algae Example\n", "Author(s): Paul Miles | Date Created: June 18, 2018\n", "\n", "This is a simplified lake algae dynamics model. We consider phytoplankton $A$, zooplankton $Z$ and nutrition $P$ (eg. phosphorus) available for $A$ in the water. The system is affected by the water outflow/inflow $Q$, incoming phosphorus load $P_{in}$ and temperature $T$. It is described as a simple predator - pray dynamics between $A$ and $Z$. The growth of $A$ is limited by the availability of $P$ and it depends on the water temperature $T$. The inflow/outflow $Q$ affects both $A$ and $P$, but not $Z$." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "# import required packages\n", "import numpy as np\n", "import scipy.io as sio\n", "from scipy.integrate import odeint\n", "from pymcmcstat.MCMC import MCMC\n", "from pymcmcstat.plotting import MCMCPlotting\n", "import matplotlib.pyplot as plt\n", "np.seterr(over='ignore');" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The data is saved in a `.mat` file as the original example comes from the Matlab. We extract the necessary data as follows" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "# load Algae data\n", "algaedata = sio.loadmat('algaedata.mat')\n", "# extract dictionary contents\n", "adata = algaedata['data']\n", "tx = adata['xdata'][0][0]\n", "ty = adata['ydata'][0][0]\n", "xlbls = adata['xlabels'][0][0][0]\n", "ylbls = adata['ylabels'][0][0][0]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Define Sum-of-Squares and Model Functions\n", "The model function requires evaluation of a system of ODEs. The solution implemented below is non-optimal. It is intended to serve as an example of how to couple and ODE system type problem with [pymcmcstat](https://prmiles.wordpress.ncsu.edu/codes/python-packages/pymcmcstat/). Optimizing your ODE system solver is left to the user's discretion." ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "def algaess(theta, data):\n", " # sum-of-squares function for algae example\n", " ndp, nbatch = data.shape[0]\n", " time = data.ydata[0][:, 0]\n", " ydata = data.ydata[0][:, 1:4]\n", " xdata = data.user_defined_object[0]\n", " # last 3 parameters are the initial states\n", " y0 = np.array(theta[-3:])\n", " # evaluate model\n", " tmodel, ymodel = algaefun(time, theta, y0, xdata)\n", " res = ymodel - ydata\n", " ss = (res**2).sum(axis=0)\n", " return ss \n", "\n", "def algaefun(time, theta, y0, xdata):\n", " \"\"\"\n", " Evaluate Ordinary Differential Equation\n", " \"\"\"\n", " sol = odeint(algaesys, y0, time, args=(theta, xdata))\n", " return time, sol\n", " \n", "def algaesys(y, t, theta, xdata):\n", " \"\"\"\n", " Model System\n", " \"\"\"\n", " A = y[0]\n", " Z = y[1]\n", " P = y[2]\n", " # control variables are assumed to be saved at each time unit interval\n", " idx = int(np.ceil(t)) - 1 \n", " if idx >= 120:\n", " idx = 119\n", " QpV = xdata[idx, 1]\n", " T = xdata[idx, 2]\n", " Pin = xdata[idx, 3]\n", " # model parameters\n", " mumax = theta[0]\n", " rhoa = theta[1]\n", " rhoz = theta[2]\n", " k = theta[3]\n", " alpha = theta[4]\n", " th = theta[5]\n", " mu = mumax*(th**(T-20))*P*((k+P)**(-1))\n", " dotA = (mu - rhoa - QpV - alpha*Z)*A\n", " dotZ = alpha*Z*A - rhoz*Z\n", " dotP = -QpV*(P-Pin) + (rhoa-mu)*A + rhoz*Z\n", " ydot = np.array([dotA, dotZ, dotP])\n", " return ydot" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Initialize MCMC Object and Setup Simulation\n", "- Define data structure\n", "- Assign parameters and define constraints\n", "- Set simulation options and model settings" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [], "source": [ "# initialize MCMC object\n", "mcstat = MCMC()\n", "# initialize data structure \n", "mcstat.data.add_data_set(x=tx[:, 0],\n", " y=ty[:, 0:4],\n", " user_defined_object=tx)\n", "# initialize parameter array\n", "#theta = [0.5, 0.03, 0.1, 10, 0.02, 1.14, 0.77, 1.3, 10]\n", "# add model parameters\n", "mcstat.parameters.add_model_parameter(name='mumax', theta0=0.5, minimum=0)\n", "mcstat.parameters.add_model_parameter(name='rhoa', theta0=0.03, minimum=0)\n", "mcstat.parameters.add_model_parameter(name='rhoz', theta0=0.1, minimum=0)\n", "mcstat.parameters.add_model_parameter(name='k', theta0=10, minimum=0)\n", "mcstat.parameters.add_model_parameter(name='alpha', theta0=0.02, minimum=0)\n", "mcstat.parameters.add_model_parameter(name='th', theta0=1.14, minimum=0,\n", " maximum=np.inf, prior_mu=0.14,\n", " prior_sigma=0.2)\n", "# initial values for the model states\n", "mcstat.parameters.add_model_parameter(name='A0', theta0=0.77, minimum=0,\n", " maximum=np.inf, prior_mu=0.77,\n", " prior_sigma=2)\n", "mcstat.parameters.add_model_parameter(name='Z0', theta0=1.3, minimum=0,\n", " maximum=np.inf, prior_mu=1.3,\n", " prior_sigma=2)\n", "mcstat.parameters.add_model_parameter(name='P0', theta0=10, minimum=0,\n", " maximum=np.inf, prior_mu=10,\n", " prior_sigma=2)\n", "\n", "# Generate options\n", "mcstat.simulation_options.define_simulation_options(\n", " nsimu=1.0e3, updatesigma=True)\n", "# Define model object:\n", "mcstat.model_settings.define_model_settings(\n", " sos_function=algaess,\n", " sigma2=0.01**2,\n", " S20=np.array([1, 1, 2]),\n", " N0=np.array([4, 4, 4]))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The code takes some time to run, so here we simply check to make sure the data structure can be processed using our sum-of-squares function. Note, we have separate sum-of-squares for each quantity of interest and there will be a separate error variance for each as well." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "ss = [ 930.44890289 521.2441601 1278.11736803]\n" ] } ], "source": [ "# check model evaluation\n", "theta = [0.5, 0.03, 0.1, 10, 0.02, 1.14, 0.77, 1.3, 10]\n", "ss = algaess(theta, mcstat.data)\n", "print('ss = {}'.format(ss))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Run simulation\n", "- We run an initialize sequence of 1000, then restart and perform another 5000" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "Sampling these parameters:\n", " name start [ min, max] N( mu, sigma^2)\n", " mumax: 0.50 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " rhoa: 0.03 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " rhoz: 0.10 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " k: 10.00 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " alpha: 0.02 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " th: 1.14 [ 0.00e+00, inf] N( 0.14, 0.20^2)\n", " A0: 0.77 [ 0.00e+00, inf] N( 0.77, 2.00^2)\n", " Z0: 1.30 [ 0.00e+00, inf] N( 1.30, 2.00^2)\n", " P0: 10.00 [ 0.00e+00, inf] N( 10.00, 2.00^2)\n", " [-----------------100%-----------------] 1001 of 1000 complete in 119.1 sec\n", "Sampling these parameters:\n", " name start [ min, max] N( mu, sigma^2)\n", " mumax: 0.39 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " rhoa: 0.03 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " rhoz: 0.11 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " k: 10.44 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " alpha: 0.02 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " th: 1.00 [ 0.00e+00, inf] N( 0.14, 0.20^2)\n", " A0: 1.30 [ 0.00e+00, inf] N( 0.77, 2.00^2)\n", " Z0: 2.09 [ 0.00e+00, inf] N( 1.30, 2.00^2)\n", " P0: 9.64 [ 0.00e+00, inf] N( 10.00, 2.00^2)\n", " [-----------------100%-----------------] 5000 of 5000 complete in 522.5 sec" ] } ], "source": [ "# Run simulation\n", "mcstat.run_simulation()\n", "# Rerun starting from results of previous run\n", "mcstat.simulation_options.nsimu = int(5.0e3)\n", "mcstat.run_simulation(use_previous_results=True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Extract results and plot chain diagnostics\n", "- chain panel\n", "- density panel\n", "- pairwise correlation panel" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "---------------------\n", "name : mean std MC_err tau geweke\n", "mumax : 0.5500 0.1332 0.0266 653.2883 0.7246\n", "rhoa : 0.0664 0.0351 0.0063 497.2180 0.1512\n", "rhoz : 0.1003 0.0059 0.0009 238.1233 0.9029\n", "k : 12.0906 3.5540 0.6679 624.3892 0.6571\n", "alpha : 0.0237 0.0013 0.0002 199.3897 0.9008\n", "th : 1.0069 0.0157 0.0024 240.8882 0.9587\n", "A0 : 1.0016 0.3543 0.0319 44.6822 0.7331\n", "Z0 : 1.8666 0.4907 0.0435 74.6226 0.8759\n", "P0 : 8.9193 0.9699 0.0972 75.9392 0.9266\n", "---------------------\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "