ies1-intro-to-compact-ga.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The pair of fruits with the highest nutritional value is: [('pear', 50), ('cashew', 50)]\n"
     ]
    }
   ],
   "source": [
    "\"\"\"Introduction to Compact Genetic Algorithm (CGA).\"\"\"\n",
    "\n",
    "# Imports\n",
    "from pyeasyga import pyeasyga\n",
    "from random import random\n",
    "\n",
    "# Fruit options\n",
    "data = [('apple', 35), ('orange', 40), ('pear', 50),\n",
    "        ('grape', 20), ('banana', 40), ('cashew', 50)]\n",
    "\n",
    "# Initialize genetic algorithm\n",
    "cga = pyeasyga.GeneticAlgorithm(data)\n",
    "\n",
    "# Update probability vector\n",
    "def update_prob(winner, loser, prob, popsize):\n",
    "  for i in range(0, len(prob)):\n",
    "    if winner[i] != loser[i]:\n",
    "      if winner[i] == 1:\n",
    "        prob[i] += 1.0 / float(popsize)\n",
    "      else:\n",
    "        prob[i] -= 1.0 / float(popsize)\n",
    "\n",
    "# Create a new individual\n",
    "def create_individual(prob):\n",
    "  individual = []\n",
    "  for p in prob:\n",
    "    if random() < p:\n",
    "      individual.append(1)\n",
    "    else:\n",
    "      individual.append(0)\n",
    "  return pyeasyga.Chromosome(individual)\n",
    "cga.create_individual = create_individual\n",
    "\n",
    "# Make competition between two individuals\n",
    "def compete(a, b):\n",
    "  if a.fitness > b.fitness:\n",
    "    return a, b\n",
    "  else:\n",
    "    return b, a\n",
    "\n",
    "# Set fitness function\n",
    "def fitness_function(individual, data):\n",
    "  fitness = 0\n",
    "  # Only pairs of fruit are valid\n",
    "  if individual.count(1) == 2:\n",
    "    # Calculate the fitness according to the fruit nutritive value\n",
    "    for (selected, (fruit, profit)) in zip(individual, data):\n",
    "      if selected:\n",
    "        fitness += profit\n",
    "  return fitness\n",
    "cga.fitness_function = fitness_function\n",
    "\n",
    "# Set evolution function\n",
    "def run(self):\n",
    "  # Initialize probability vector\n",
    "  prob = [0.5] * len(data)\n",
    "  # Initialize best solution\n",
    "  best = None\n",
    "  # Run i generations\n",
    "  for _ in range(0, self.generations):\n",
    "    # Create individuals\n",
    "    a = self.create_individual(prob)\n",
    "    b = self.create_individual(prob)\n",
    "    # Calculate fitness for each individual\n",
    "    a.fitness = self.fitness_function(a.genes, self.seed_data)\n",
    "    b.fitness = self.fitness_function(b.genes, self.seed_data)\n",
    "    # Get the best and worst individual\n",
    "    winner, loser = compete(a, b)\n",
    "    # Update best solution\n",
    "    if best:\n",
    "      if winner.fitness > best.fitness:\n",
    "        best = winner\n",
    "    else:\n",
    "      best = winner\n",
    "    # Update the probability vector based on the success of each bit\n",
    "    update_prob(winner.genes, loser.genes, prob, self.population_size)\n",
    "  # Add final solution\n",
    "  self.current_generation.append(best)\n",
    "cga.run = run\n",
    "\n",
    "# Run cGA\n",
    "cga.run(cga)\n",
    "# Get best individual\n",
    "result = cga.best_individual()\n",
    "\n",
    "# Prepare result to be printed\n",
    "def prepare_result(result):\n",
    "  profit, individual = result\n",
    "  pair = []\n",
    "  for i, gene in enumerate(individual):\n",
    "    if gene == 1:\n",
    "      pair.append(data[i])\n",
    "  return pair\n",
    "\n",
    "# Print result\n",
    "print('The pair of fruits with the highest nutritional value is: {}'\n",
    "  .format(prepare_result(result)))"
   ]
  }
 ],
 "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}