ES_a2c_algorithm.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Sources =>\n",
    "ES reference \n",
    "https://github.com/alirezamika/evostra/blob/master/evostra/algorithms/evolution_strategy.py\n",
    " \n",
    "A2C reference\n",
    "https://github.com/MorvanZhou/Reinforcement-learning-with-tensorflow/\n",
    "'''\n",
    "import tensorflow as tf\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.optimizers import Adam\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import random\n",
    "import gym\n",
    "import math\n",
    "import matplotlib.pyplot as plt\n",
    "from gym.wrappers.monitor import Monitor\n",
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "device_name = \"/gpu:0\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Softmax for fitness \n",
    "'''\n",
    "def softmax(x):\n",
    "    ex = np.exp(x-np.max(x))\n",
    "    return ex/ex.sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Rank population based on fitness score\n",
    "'''\n",
    "def fitness(env, w0, w1):\n",
    "    total_reward = 0.0\n",
    "    num_run = 1\n",
    "    \n",
    "    # calculate weight that is expected to give maximum reward \n",
    "    # over time\n",
    "    for t in range(num_run):\n",
    "        observation = env.reset()\n",
    "        for i in range(200):\n",
    "            p0 = np.dot(w0, observation)\n",
    "            p1 = np.dot(w1, observation)\n",
    "            joint = [p0, p1]\n",
    "            softmax_p = softmax(joint)\n",
    "            \n",
    "            #print (softmax_p)\n",
    "            action = 0 if random.uniform(0,1) < softmax_p[0] else 1\n",
    "            observation, reward, done, info = env.step(action)\n",
    "\n",
    "            total_reward += reward\n",
    "            if done:\n",
    "                break\n",
    "    return total_reward / num_run"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Tournament selection \n",
    "'''\n",
    "def evolution_strategy(env, weight, policy_grad, value_grad, sess):\n",
    "    # sample from population\n",
    "    population_size = 50 \n",
    "    \n",
    "    # jitter weights using noise std dev\n",
    "    noise = 0.1 \n",
    "    \n",
    "    # fitness fun\n",
    "    reward = fitness(env, weight[:,0], weight[:,1])\n",
    "    \n",
    "    # sample from normal distribution\n",
    "    N = np.random.randn(population_size, 4,2) \n",
    "    R = np.zeros(population_size)\n",
    "    \n",
    "    maxReward = 0\n",
    "    maxWeight = weight\n",
    "    \n",
    "    with tf.device(device_name):\n",
    "        for j in range(population_size):\n",
    "            \n",
    "            # jitter the weight using the gaussian\n",
    "            noise_weight = weight + noise * N[j,:,:]\n",
    "            assign_params(policy_grad, sess, noise_weight)\n",
    "\n",
    "            for i in range(5):\n",
    "                run_episode(env, policy_grad, value_grad, sess)\n",
    "\n",
    "            with tf.variable_scope(\"policy\", reuse = tf.AUTO_REUSE):\n",
    "                weight_tf = tf.get_variable(\"policy_parameters\",[4,2], dtype = tf.float64)\n",
    "\n",
    "            resWeights = weight_tf.eval()\n",
    "            weight0 = resWeights[:,0]\n",
    "            weight1 = resWeights[:,1]\n",
    "            \n",
    "            # get reward from the jittered version\n",
    "            R[j] = fitness(env, weight0, weight1) \n",
    "\n",
    "            if R[j] > maxReward:\n",
    "                maxReward = R[j]\n",
    "                maxWeight = resWeights\n",
    "\n",
    "    return maxWeight, reward"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Policy estimation\n",
    "'''\n",
    "def policyGradient():\n",
    "    with tf.variable_scope(\"policy\", reuse = tf.AUTO_REUSE):\n",
    "        params = tf.get_variable(\"policy_parameters\",[4,2], dtype = tf.float64)\n",
    "        \n",
    "        c = tf.placeholder(\"float64\", [4, 2])\n",
    "        s = tf.placeholder(\"float64\", [None, 4])\n",
    "        a = tf.placeholder(\"float64\", [None, 2])\n",
    "        adv = tf.placeholder(\"float64\", [None, 1])\n",
    "        \n",
    "        with tf.device(device_name):\n",
    "            #linear layer\n",
    "            linear = tf.matmul(s, params)\n",
    "            p = tf.nn.softmax(linear)\n",
    "            \n",
    "            #print p.get_shape()\n",
    "            \n",
    "            p_res = tf.reduce_sum(tf.multiply(p, a), reduction_indices=[1])\n",
    "            eligibility = tf.log(p_res) * adv\n",
    "            \n",
    "            #print p_res.get_shape()\n",
    "            \n",
    "            \n",
    "            loss = -tf.reduce_sum(eligibility)\n",
    "            optimizer = tf.train.AdamOptimizer(0.01).minimize(loss)\n",
    "\n",
    "            params_assign = params.assign(c)\n",
    "\n",
    "        return p, s, a, adv, optimizer, c, params_assign"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Gradient estimation\n",
    "'''\n",
    "def valueGradient():\n",
    "    with tf.variable_scope(\"value\"):\n",
    "        with tf.device(device_name):\n",
    "            state = tf.placeholder(\"float\",[None,4])\n",
    "            res = tf.placeholder(\"float\",[None,1])\n",
    "            \n",
    "            #l1\n",
    "            w1 = tf.get_variable(\"w1\",[4,10])\n",
    "            b1 = tf.get_variable(\"b1\",[10])\n",
    "            \n",
    "            #hidden\n",
    "            h1 = tf.nn.relu(tf.matmul(state,w1) + b1)\n",
    "            \n",
    "            #l2\n",
    "            w2 = tf.get_variable(\"w2\",[10,1])\n",
    "            b2 = tf.get_variable(\"b2\",[1])\n",
    "\n",
    "            out = tf.matmul(h1, w2) + b2\n",
    "            error = out - res\n",
    "            loss = tf.nn.l2_loss(error)\n",
    "\n",
    "            optimizer = tf.train.AdamOptimizer(0.1).minimize(loss)\n",
    "        return out, state, res, optimizer, loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def assign_params(policy_grad, sess, w_try):\n",
    "    pl_calculated, pl_state, pl_actions, pl_advantages, pl_optimizer, c, params_assign = policy_grad\n",
    "    sess.run(params_assign, feed_dict={c: w_try})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_episode(env, policyGrad, valueGrad, sess):\n",
    "    pl_calculated, pl_state, pl_actions, pl_advantages, pl_optimizer, c, params_assign = policyGrad\n",
    "    vl_calculated, vl_state, vl_newvals, vl_optimizer, vl_loss = valueGrad\n",
    "    \n",
    "    observation = env.reset()\n",
    "    total_reward = 0\n",
    "    \n",
    "    states = []\n",
    "    actions = []\n",
    "    advantages = []\n",
    "    transitions = []\n",
    "    updatedValues = []\n",
    "\n",
    "    for _ in range(200):\n",
    "\n",
    "        # get the policy \n",
    "        observationList = np.expand_dims(observation, axis=0)\n",
    "        probs = sess.run(pl_calculated,feed_dict={pl_state: observationList})\n",
    "        action = 0 if random.uniform(0,1) < probs[0][0] else 1\n",
    "\n",
    "        # get transition\n",
    "        states.append(observation)\n",
    "        actionChoice = np.zeros(2)\n",
    "        actionChoice[action] = 1\n",
    "        actions.append(actionChoice)\n",
    "\n",
    "        # move in the environment\n",
    "        previousObservations = observation\n",
    "        observation, reward, done, info = env.step(action)\n",
    "\n",
    "        transitions.append((previousObservations, action, reward))\n",
    "        total_reward += reward\n",
    "\n",
    "        if done:\n",
    "            break\n",
    "    \n",
    "    for i, t in enumerate(transitions):\n",
    "        obs, action, reward = t\n",
    "\n",
    "        # get discounted monte-carlo return \n",
    "        newReward = 0\n",
    "        newTransitions = len(transitions) - i\n",
    "        decrease = 1\n",
    "        \n",
    "        for j in range(newTransitions):\n",
    "            newReward += transitions[(j) + i][2] * decrease\n",
    "            decrease = decrease * 0.97\n",
    "        \n",
    "        observationList = np.expand_dims(obs, axis=0)\n",
    "        currentval = sess.run(vl_calculated,feed_dict={vl_state: observationList})[0][0]\n",
    "\n",
    "        advantages.append(newReward - currentval)\n",
    "        updatedValues.append(newReward)\n",
    "\n",
    "    # save state for updated \n",
    "    updatedValuesList = np.expand_dims(updatedValues, axis=1)\n",
    "    sess.run(vl_optimizer, feed_dict={vl_state: states, vl_newvals: updatedValuesList})\n",
    "\n",
    "    advantagesList = np.expand_dims(advantages, axis=1)\n",
    "    sess.run(pl_optimizer, feed_dict={pl_state: states, pl_advantages: advantagesList, pl_actions: actions})\n",
    "    \n",
    "    return total_reward"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.6/site-packages/tensorflow/python/client/session.py:1702: UserWarning: An interactive session is already active. This can cause out-of-memory errors in some cases. You must explicitly call `InteractiveSession.close()` to release resources held by the other session(s).\n",
      "  warnings.warn('An interactive session is already active. This can '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 10.0\n",
      "1 12.0\n",
      "2 43.0\n",
      "3 16.0\n",
      "4 22.0\n",
      "5 9.0\n",
      "6 15.0\n",
      "7 15.0\n",
      "8 35.0\n",
      "9 20.0\n",
      "10 30.0\n",
      "11 13.0\n",
      "12 24.0\n",
      "13 35.0\n",
      "14 35.0\n",
      "15 29.0\n",
      "16 11.0\n",
      "17 39.0\n",
      "18 35.0\n",
      "19 18.0\n",
      "20 35.0\n",
      "21 103.0\n",
      "22 38.0\n",
      "23 64.0\n",
      "24 17.0\n",
      "25 116.0\n",
      "26 57.0\n",
      "27 88.0\n",
      "28 25.0\n",
      "29 28.0\n",
      "30 38.0\n",
      "31 95.0\n",
      "32 92.0\n",
      "33 141.0\n",
      "34 110.0\n",
      "35 46.0\n",
      "36 70.0\n",
      "37 199.0\n",
      "38 168.0\n",
      "39 35.0\n",
      "40 76.0\n",
      "41 134.0\n",
      "42 56.0\n",
      "43 117.0\n",
      "44 200.0\n",
      "45 145.0\n",
      "46 200.0\n",
      "47 52.0\n",
      "48 60.0\n",
      "49 28.0\n",
      "50 36.0\n",
      "51 200.0\n",
      "52 157.0\n",
      "53 124.0\n",
      "54 109.0\n",
      "55 100.0\n",
      "56 83.0\n",
      "57 200.0\n",
      "58 84.0\n",
      "59 147.0\n",
      "60 200.0\n",
      "61 200.0\n",
      "62 193.0\n",
      "63 181.0\n",
      "64 127.0\n",
      "65 48.0\n",
      "66 191.0\n",
      "67 200.0\n",
      "68 200.0\n",
      "69 148.0\n",
      "70 87.0\n",
      "71 174.0\n",
      "72 200.0\n",
      "73 186.0\n",
      "74 147.0\n",
      "75 200.0\n",
      "76 184.0\n",
      "77 200.0\n",
      "78 200.0\n",
      "79 200.0\n",
      "80 200.0\n",
      "81 118.0\n",
      "82 200.0\n",
      "83 115.0\n",
      "84 146.0\n",
      "85 200.0\n",
      "86 200.0\n",
      "87 200.0\n",
      "88 200.0\n",
      "89 200.0\n",
      "90 123.0\n",
      "91 200.0\n",
      "92 200.0\n",
      "93 200.0\n",
      "94 200.0\n",
      "95 200.0\n",
      "96 149.0\n",
      "97 200.0\n",
      "98 200.0\n",
      "99 200.0\n",
      "100 200.0\n",
      "101 156.0\n",
      "102 193.0\n",
      "103 184.0\n",
      "104 200.0\n",
      "105 200.0\n",
      "106 200.0\n",
      "107 200.0\n",
      "108 158.0\n",
      "109 143.0\n",
      "110 200.0\n",
      "111 200.0\n",
      "112 130.0\n",
      "113 158.0\n",
      "114 172.0\n",
      "115 106.0\n",
      "116 200.0\n",
      "117 200.0\n",
      "118 200.0\n",
      "119 200.0\n",
      "120 200.0\n",
      "121 200.0\n",
      "122 188.0\n",
      "123 196.0\n",
      "124 159.0\n",
      "125 194.0\n",
      "126 200.0\n",
      "127 200.0\n",
      "128 200.0\n",
      "129 114.0\n",
      "130 200.0\n",
      "131 128.0\n",
      "132 125.0\n",
      "133 158.0\n",
      "134 129.0\n",
      "135 121.0\n",
      "136 200.0\n",
      "137 200.0\n",
      "138 200.0\n",
      "139 165.0\n",
      "140 200.0\n",
      "141 134.0\n",
      "142 200.0\n",
      "143 200.0\n",
      "144 155.0\n",
      "145 200.0\n",
      "146 175.0\n",
      "147 200.0\n",
      "148 200.0\n",
      "149 120.0\n",
      "150 200.0\n",
      "151 200.0\n",
      "152 200.0\n",
      "153 200.0\n",
      "154 200.0\n",
      "155 200.0\n",
      "156 200.0\n",
      "157 200.0\n",
      "158 148.0\n",
      "159 200.0\n",
      "160 141.0\n",
      "161 191.0\n",
      "162 200.0\n",
      "163 200.0\n",
      "164 200.0\n",
      "165 165.0\n",
      "166 200.0\n",
      "167 200.0\n",
      "168 113.0\n",
      "169 199.0\n",
      "170 200.0\n",
      "171 127.0\n",
      "172 71.0\n",
      "173 200.0\n",
      "174 200.0\n",
      "175 187.0\n",
      "176 200.0\n",
      "177 137.0\n",
      "178 200.0\n",
      "179 200.0\n",
      "180 149.0\n",
      "181 200.0\n",
      "182 200.0\n",
      "183 138.0\n",
      "184 200.0\n",
      "185 200.0\n",
      "186 179.0\n",
      "187 200.0\n",
      "188 135.0\n",
      "189 200.0\n",
      "190 200.0\n",
      "191 200.0\n",
      "192 131.0\n",
      "193 170.0\n",
      "194 103.0\n",
      "195 200.0\n",
      "196 200.0\n",
      "197 154.0\n",
      "198 124.0\n",
      "199 131.0\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe63819a240>]"
      ]
     },
     "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": [
    "'''\n",
    "Test network on cartpole\n",
    "'''\n",
    "\n",
    "cartpoleRewards = np.zeros(200)\n",
    "\n",
    "for i in range(1):\n",
    "    env = gym.make('CartPole-v0')\n",
    "    monitor = Monitor(env,'cartpole-hill/', force=True)\n",
    "    tf.reset_default_graph() \n",
    "    \n",
    "    pG = policyGradient()\n",
    "    vG = valueGradient()\n",
    "    \n",
    "    sess = tf.InteractiveSession()\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    \n",
    "    with tf.variable_scope(\"policy\", reuse = tf.AUTO_REUSE):\n",
    "        weight = tf.get_variable(\"policy_parameters\", [4,2], dtype = tf.float64).eval()\n",
    "    \n",
    "    for i in range(200):\n",
    "        weight, reward = evolution_strategy(env, weight, pG, vG, sess)\n",
    "        print(i, reward)\n",
    "\n",
    "        cartpoleRewards[i] = reward\n",
    "\n",
    "    sess.close()\n",
    "    monitor.close()\n",
    "    env.close()\n",
    "\n",
    "plt.title(\"Evolutionary A2C\")\n",
    "plt.plot(cartpoleRewards)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe63b61d9e8>]"
      ]
     },
     "execution_count": 21,
     "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.title(\"Evolutionary A2C\")\n",
    "plt.plot(cartpoleRewards)"
   ]
  },
  {
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}