mas_socialWelfare_basics_fr.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# La théorie du bien-être social\n",
    "\n",
    "Auteurs : Philippe Mathieu et Antoine Nongaillard, [CRISTAL Lab](https://www.cristal.univ-lille.fr/), [SMAC Team](https://www.cristal.univ-lille.fr/?rubrique26&id=7), [University of Lille](http://www.univ-lille1.fr), email : philippe.mathieu@univ-lille.fr\n",
    "\n",
    "Creation : 18/12/2021\n",
    "\n",
    "\n",
    "## Principe\n",
    "\n",
    "Le problème présenté ici, connu sous l'acronyme **MARA** (MultiAgent Resource Allocation), rentre dans le cadre de l'affectation de ressources à un certain nombre d'agents. On considère un ensemble de ressources et un ensemble d'agents à qui l'on doit distribuer ces ressources. Les ressources considérées ici sont discrètes, indivisibles et non partageables. Chaque agent étant autonome, il n'accorde pas la même importance aux différentes ressources que ses congénères : il exprime ses préférences au moyen d'une valeur d'utilité associée à chaque ressource. Pour mesurer sa satisfaction individuelle, on se base sur l'agrégation des différentes valeurs associées aux ressources que possède un agent. D'autres considérations pourraient bien-sûr être prises en compte, nous restons ici dans un cadre simple. Le bien-être individuel d'un agent est évalué par la somme des utilités qu'il accorde aux ressources qu'il possède (on dit que les fonctions d'utilité sont additives).\n",
    "\n",
    "La collectivité doit aussi s'accorder sur une mesure de bien-être. Pour mesurer le bien-être social, il existe 4 grandes méthodes classiques :\n",
    "\n",
    "- Utilitaire : maximiser la somme des utilités des agents\n",
    "- Égalitaire : maximiser le minimum des utilités des agents\n",
    "- Élitiste : maximiser le maximum des utilités des agents\n",
    "- Nash : maximiser le produit des utilités des agents\n",
    "\n",
    "Le problème consiste alors à trouver comment affecter l'ensemble des ressources aux différents agents de manière à maximiser le bien-être social choisi.\n",
    "\n",
    "Cette feuille montre comment, à l'aide d'une approche comportementale, il est possible de s'approcher, voire de calculer, l'optimum social.\n",
    "\n",
    "### Comment représenter ressources et fonctions d'utilité\n",
    "\n",
    "Afin de pouvoir jouer sur les nombres de ressources et/ou d'agents, nous générerons aléatoirement des noms de ressources afin de pouvoir en avoir autant que l'on souhaite. Pour la fonction d'utilité, on utilisera un dictionnaire Python, dont les clés sont les ressources potentielles et les valeurs le poids que l'agent y accorde. Pour ces valeurs on prendra un entier naturel (donc sans le zéro, qui évidemment pose problème lors d'un produit de Nash). Par convention, plus une ressource est importante aux yeux d'un agent, plus la valeur qu'il lui associe est grande.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Un ensemble de ressources \t ['DAGOY', 'TPTJD', 'GVDTB', 'CYUFE', 'TURHV', 'RCHLX', 'NUCRD', 'XERVK', 'ULFCW', 'PFYDI']\n",
      "Une fonction d'utilité\t {'DAGOY': 16, 'TPTJD': 7, 'GVDTB': 9, 'CYUFE': 16, 'TURHV': 18, 'RCHLX': 5, 'NUCRD': 16, 'XERVK': 9, 'ULFCW': 16, 'PFYDI': 2}\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd # mainly for results\n",
    "import random       # mainly for choice et randrange\n",
    "# import string     # mainly for ascii_uppercase\n",
    "import functools    # mainly for reduce\n",
    "import copy         # mainly for deepcopy \n",
    "\n",
    "nb_resources =10             # nombre de ressources souhaités\n",
    "valeurs = list(range(1,20))  # intervalle de valuation de chaque ressource\n",
    "\n",
    "def genere_nom():\n",
    "    return ''.join(random.choice('ABCDEFGHIJKLMNOPQRSTUVWXYZ') for _ in range(5))\n",
    "\n",
    "# Générer une liste de ressources\n",
    "ressources = [genere_nom() for _ in range(nb_resources)]\n",
    "print(\"Un ensemble de ressources \\t\", ressources)\n",
    "\n",
    "# Créer une fonction d'utilité avec des valeurs aléatoires\n",
    "u = {r : random.choice(valeurs) for r in ressources}\n",
    "print(\"Une fonction d'utilité\\t\" , u)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## On passe aux agents\n",
    "Un agent est défini avec le lot (`bag`) de ressources qu'il possède et une fonction d'utilité `fUtil` lui permettant de l'évaluer.\n",
    "Lors de sa création, la fonction d'évaluation de l'agent est définie aléatoirement. Le lot est initialement vide. La méthode `individual_welfare` permet de connaître le bien-être individuel de l'agent, ou en d'autres termes, comment l'agent évalue son lot."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Agent:\n",
    "    def __init__(self,name) :\n",
    "        self.name=name\n",
    "        self.fUtil = {r : random.choice(valeurs) for r in ressources}\n",
    "        self.bag = []\n",
    "    def __str__(self) :\n",
    "        return \"agent \"+str(self.name)+\"  \"+ str(len(self.bag)) + \" resources. Individual welfare \"+ str(self.individual_welfare())\n",
    "    def individual_welfare(self):\n",
    "        return sum([self.fUtil.get(r) for r in self.bag]) \n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Quelques manipulations basiques sur un agent pour s'entrainer ;-)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Bag \t\t ['DAGOY', 'NUCRD', 'CYUFE', 'XERVK', 'TPTJD']\n",
      "fUtil \t\t {'DAGOY': 17, 'TPTJD': 4, 'GVDTB': 3, 'CYUFE': 13, 'TURHV': 14, 'RCHLX': 4, 'NUCRD': 14, 'XERVK': 18, 'ULFCW': 5, 'PFYDI': 15}\n",
      "individual welfare \t 66\n",
      "valued bag \t [('DAGOY', 17), ('NUCRD', 14), ('CYUFE', 13), ('XERVK', 18), ('TPTJD', 4)]\n",
      "valued and sorted bag \t [('TPTJD', 4), ('CYUFE', 13), ('NUCRD', 14), ('DAGOY', 17), ('XERVK', 18)]\n",
      "sorted resources\t ['TPTJD', 'CYUFE', 'NUCRD', 'DAGOY', 'XERVK']\n"
     ]
    }
   ],
   "source": [
    "a=Agent(\"ag\")\n",
    "a.bag=random.sample(ressources,5) # choices -> avec remplacement ; sample -> sans remplacement\n",
    "\n",
    "# le bag de l'agent a\n",
    "print(\"Bag \\t\\t\", a.bag)                              # ou sorted(a.bag)\n",
    "\n",
    "# la fonction d'utilité de l'agent a\n",
    "print(\"fUtil \\t\\t\", a.fUtil)                        # ou dict(sorted(a.fUtil.items()))\n",
    "\n",
    "# le bien-être de l'agent a\n",
    "print(\"individual welfare \\t\", a.individual_welfare())\n",
    "\n",
    "# le bag avec ses valeurs\n",
    "bv = [ (r,a.fUtil.get(r)) for r in a.bag]\n",
    "print(\"valued bag \\t\" , bv )\n",
    "\n",
    "# le bag valué , trié par ordre croissant des valeurs\n",
    "print(\"valued and sorted bag \\t\", sorted(bv, key=lambda x: x[1]) )\n",
    "\n",
    "# les ressources, triées par valuation (la premier est celle dont on se débarrasse le plus facilement\n",
    "print(\"sorted resources\\t\", [r for r,_ in sorted(bv, key=lambda x: x[1])]    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On peut maintenant créer autant d'agents que l'on souhaite, leur affecter des ressources et calculer leur bien-être individuel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "agent ag0  2 resources. Individual welfare 12\n",
      "agent ag1  1 resources. Individual welfare 19\n",
      "agent ag2  3 resources. Individual welfare 9\n",
      "agent ag3  2 resources. Individual welfare 23\n",
      "agent ag4  2 resources. Individual welfare 27\n"
     ]
    }
   ],
   "source": [
    "# créer 5 agents\n",
    "agentList = [Agent(\"ag\"+str(i)) for i in range(5)]\n",
    "\n",
    "# distribuer les ressources aléatoirement entre les agents\n",
    "for i in range(len(ressources)) :\n",
    "    random.choice(agentList).bag.append(ressources[i])\n",
    "\n",
    "# afficher les agents\n",
    "for a in agentList: print(a)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Approche centralisée\n",
    "Maintenant que les agents ont un `bag` rempli, il est maintenant possible de calculer le bien-être de cette société d'agents.\n",
    "Le calcul du bien-être social consiste à agréger (selon la fonction de bien-être souhaitée) les bien-êtres individuels de chaque agent. Par exemple, le bien-être-social utilitaire correspond à la somme des bien-êtres individuels.\n",
    "\n",
    "| Nom | fonction à maximiser | Caractéristiques |\n",
    "| -----| ----|--- |\n",
    "| utilitaire | somme | équilibre global, sans tenir compte des bien-être individuels |\n",
    "| égalitaire | min | permet de satisfaire un minimum de désir pour tous les agents \n",
    "| élitiste | max | on donne tout à son \"champion\" |\n",
    "| Nash | produit | réduire les écarts entre les individus, diminution des inégalités |\n",
    "\n",
    "Notons par ailleurs que la comparaison des valeurs atteintes par ces différentes fonctions de bien-être social n'a pas de sens."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Utilitarian Social Welfare\t 90\n",
      "Egalitarian Social Welfare\t 9\n",
      "Elitist Social Welfare\t\t 27\n",
      "Nash Social Welfare\t\t 1274292\n"
     ]
    }
   ],
   "source": [
    "print('Utilitarian Social Welfare\\t', functools.reduce(lambda a, b: sum([a,b]) , [a.individual_welfare() for a in agentList]) )\n",
    "print('Egalitarian Social Welfare\\t', functools.reduce(lambda a, b: min([a,b]) , [a.individual_welfare() for a in agentList]) )\n",
    "print('Elitist Social Welfare\\t\\t',   functools.reduce(lambda a, b: max([a,b]) , [a.individual_welfare() for a in agentList]) )\n",
    "print('Nash Social Welfare\\t\\t', functools.reduce(lambda a, b: np.prod([a,b]) , [a.individual_welfare() for a in agentList]) )\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Notion d'optimal social\n",
    "\n",
    "Parmi toutes les affectations possibles des différentes ressources, certaines fournissent la valeur maximale possible à la fonction de bien-être social. Pour certains bien-être sociaux, cette affectation est facile à calculer (bien-être élitiste ou utilitaire) alors que pour d'autres, c'est beaucoup plus compliqué. De plus, on considère ici qu'une \"main invisible\" (clin d'œil à [Adam Smith](https://fr.wikipedia.org/wiki/Main_invisible)) distribue les ressources comme par magie. Si on souhaite prendre en compte les relations sociales entre les agents, cela devient compliqué même pour les fonctions d'utilité les plus simples.\n",
    "\n",
    "##### Solution optimale pour le bien-être social élitiste\n",
    "Le bien-être social **élitiste** consiste à **maximiser la richesse du plus riche** des agents (tout donner à son \"champion\" en quelque sorte). Quand les utilités sont additives, calculer cet optimal est facile puisqu'il suffit de donner toutes les ressources à l'agent qui valorise le mieux la totalité des ressources.\n",
    "La \"main invisible\" affecte tour à tour toutes les ressources à chaque agent et regarde lequel est associé à la valeur de bien-être individuel maximum (on pourrait imaginer faire la somme des valuations de la fonction d'utilité sans affecter les ressources, mais cela ne fonctionnerait que si toutes les ressources sont uniques)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elitist. Optimal value :  113 . Obtained by giving all the resources to agent  ag0\n"
     ]
    }
   ],
   "source": [
    "# on affecte toutes les ressources à tous les agents\n",
    "for a in agentList:\n",
    "    a.bag = ressources.copy()\n",
    "\n",
    "# On recherche la plus forte évaluation\n",
    "index = np.argmax([a.individual_welfare() for a in agentList])\n",
    "m = agentList[index].individual_welfare()\n",
    "\n",
    "print(\"Elitist. Optimal value : \", m,\". Obtained by giving all the resources to agent \", agentList[index].name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Solution optimale pour le bien-être social utilitaire\n",
    "Le bien-être social **utilitaire** consiste à **maximiser la somme** des richesses de tous les agents. Calculer cet optimal est facile puisqu'il suffit de donner chaque ressource à l'agent qui la valorise le mieux."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Utilitarist. Optimal value :  164  obtained using the following affectations :\n",
      "agent ag0  2 resources. Individual welfare 36 ['DAGOY', 'TURHV']\n",
      "agent ag1  2 resources. Individual welfare 33 ['TPTJD', 'PFYDI']\n",
      "agent ag2  2 resources. Individual welfare 32 ['CYUFE', 'XERVK']\n",
      "agent ag3  2 resources. Individual welfare 37 ['GVDTB', 'NUCRD']\n",
      "agent ag4  2 resources. Individual welfare 26 ['RCHLX', 'ULFCW']\n"
     ]
    }
   ],
   "source": [
    "# on réinitialise les bags\n",
    "for a in agentList :\n",
    "        a.bag=[]\n",
    "\n",
    "# on affecte les ressources 1 par 1 à l'agent qui la valorise le mieux               \n",
    "for r in ressources :\n",
    "        index = np.argmax([a.fUtil.get(r)  for a in agentList])\n",
    "        ag = agentList[index]\n",
    "        ag.bag.append(r)\n",
    "\n",
    "# On fait la somme des évaluation\n",
    "s = sum([a.individual_welfare() for a in agentList])\n",
    "\n",
    "print(\"Utilitarist. Optimal value : \", s ,\" obtained using the following affectations :\")\n",
    "for a in agentList :\n",
    "        print(a, a.bag)   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Solution optimale pour les bien-êtres égalitaires ou Nash\n",
    "Cette fois, il n'y a pas de solution simple garantissant l'optimum, même pour une \"main invisible\". Il faudrait tester toutes les allocations possibles. Avec `n` agents et `r` ressources il existe `n^r` affectations possibles (si on a 10 agents et 100 ressources cela fait 10^100 tests, ce qui est hors de portée des ordinateurs actuels).\n",
    "On peut alors éventuellement utiliser des méthodes approximatives type _Monte-Carlo_ ou _Algorithme génétique_, mais rien ne garantit l'optimalité du résultat.\n",
    "\n",
    "Ci-dessous une méthode de _Monte-Carlo_ pour le bien-être de Nash ..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best allocation 27648192 {'ag0': ['DAGOY', 'GVDTB', 'TURHV'], 'ag1': ['TPTJD', 'PFYDI'], 'ag2': ['CYUFE', 'XERVK'], 'ag3': ['NUCRD'], 'ag4': ['RCHLX', 'ULFCW']}\n"
     ]
    }
   ],
   "source": [
    "bestWelfare=-1\n",
    "bestAlloc=[]\n",
    "for i in range(100000):\n",
    "    # on réinitialise les bags\n",
    "    for a in agentList :\n",
    "        a.bag=[]\n",
    "    # on affecte les ressources au hasard\n",
    "    for r in ressources :\n",
    "        a = random.choice(agentList)\n",
    "        a.bag.append(r)\n",
    "    # On compare le résultat à la meilleure situation connue\n",
    "    p = np.prod([a.individual_welfare() for a in agentList])  # sum,prod,min,max\n",
    "    if (p > bestWelfare):\n",
    "        bestWelfare=p\n",
    "        bestAlloc = {a.name:a.bag for a in agentList}\n",
    "        \n",
    "print(\"Best allocation\",bestWelfare,bestAlloc)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "... Une amélioration, basée sur les permutations de ressources, est possible puisque l'on sait dès le départ que pour Nash il est préférable d'équilibrer les ressources entre les agents."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best allocation 36571392 {'ag0': array(['DAGOY', 'TURHV'], dtype='<U5'), 'ag1': array(['TPTJD', 'PFYDI'], dtype='<U5'), 'ag2': array(['XERVK', 'CYUFE'], dtype='<U5'), 'ag3': array(['NUCRD', 'GVDTB'], dtype='<U5'), 'ag4': array(['RCHLX', 'ULFCW'], dtype='<U5')}\n"
     ]
    }
   ],
   "source": [
    "bestWelfare=-1\n",
    "bestAlloc=[]\n",
    "for i in range(100000):\n",
    "    # on réinitialise les bags\n",
    "    for a in agentList :\n",
    "        a.bag=[]\n",
    "    # on découpe les ressources en paquets égaux\n",
    "    random.shuffle(ressources)\n",
    "    distrib = np.array_split(ressources, len(agentList))\n",
    "    for i in range(len(agentList)) :\n",
    "        agentList[i].bag = distrib[i]\n",
    "    # On compare le résultat à la meilleure situation connue\n",
    "    p = np.prod([a.individual_welfare() for a in agentList])  # sum,prod,min,max\n",
    "    if (p > bestWelfare):\n",
    "        bestWelfare=p\n",
    "        bestAlloc = {a.name:a.bag for a in agentList}\n",
    "        \n",
    "print(\"Best allocation\",bestWelfare,bestAlloc)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On note ici la différence entre \"Individual-based system\" and \"agent-based system\". Jusqu'à présent il y avait bien des agents, chacun avec ses propres connaissances, mais aucun comportement. Il faut la \"main invisible\" pour résoudre le problème. La majorité du travail algorithmique se fait hors des agents. L'approche \"agent-based\" est une approche comportementale : tout se passe dans le comportement de l'agent, c'est l'agent qui décide ! Il n'y a plus de \"main invisible\". Seul un système multi-agents (SMA), qui n'est là que pour le strict minimum : donner la parole aux agents.\n",
    "\n",
    "XXXX DERNIÈRE PHRASE N'A PAS DE SENS\n",
    "\n",
    "## L'approche multi-agents\n",
    "\n",
    "L'approche centralisée présente un certain nombre de défauts :\n",
    "- Avoir une affectation optimale par cette technique n'est pas toujours possible (voir ci-dessus).\n",
    "- Cette technique oblige tous les agents à indiquer à l'organe central de calcul toutes leurs utilités. La garantie d'une certaine privacité est pourtant parfois nécessaire.\n",
    "- La \"main invisible\" ! L'approche centralisée prend difficilement en compte le fait que les agents ne se connaissent pas forcément tous. Dans la vraie vie, les connaissances des individus dessinent un graphe social.\n",
    "\n",
    "Par opposition, l'approche multi-agents possède plusieurs avantages :\n",
    "- Une approche multi-agents peut être implémentée avec une programmation distribuée. Dans ce cas, les communications sont asynchrones par envoi de message avec accusés de réception et deadline pour la robustesse. C'est via ces communications qu'un agent peut être informé de l'état du système ou plus particulièrement de l'état de ses accointances.\n",
    "- Dans une exécution distribuée, des échanges de ressources peuvent avoir lieu de manière concurrente entre différents couples d'agents.\n",
    "- L'approche distribuée est évidemment bien plus conforme aux situations réelles.\n",
    "\n",
    "Nous étudions par la suite une approche comportementale qui permet aux agents d'échanger des ressources pour améliorer le bien-être social. Chaque agent tour à tour peut effectuer certains échanges avec ces accointances fidèles à un type de réseau social. On peut alors voir peu à peu le bien-être social évoluer. L'important maintenant, n'est plus \"quel est l'optimum ?\", mais \"comment on arrive à l'optimum ?\".\n",
    "\n",
    "\n",
    "##### Tout d'abord réalisons le SMA, le réseau social et l'agent ... \n",
    "- Cette fois, c'est le SMA qui crée les agents, dans son constructeur.\n",
    "- Les agents sont initialement créés vides.\n",
    "- Des méthodes permettent d'affecter aux agents des accointances ou des ressources. Ceci permettra par la suite de relancer le même SMA soit avec la même affectation de ressources, soit avec le même réseau social, soit les deux, afin de reproduire des simulations."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Le réseau social\n",
    "\n",
    "Comme dans [mas_basics](), on utilisera la librairie [networkx](https://networkx.org/) qui contient de nombreuses fonctions de création de graphes et permet de récupérer avec sa fonction `to_numpy_array` une matrice d'adjacence."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import networkx as nx\n",
    "\n",
    "nb_nodes = 20\n",
    "graph = nx.barabasi_albert_graph(nb_nodes, 2)\n",
    "#graph = nx.complete_graph(nb_nodes)\n",
    "#graph = nx.erdos_renyi_graph(nb_nodes,0.6)\n",
    "#graph = nx.cycle_graph(nb_nodes)\n",
    "#graph = nx.random_regular_graph(3,nb_nodes)\n",
    "matrix = nx.to_numpy_array(graph, nodelist=range(nb_nodes))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Et pour la représentation graphique ..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.figure(figsize=(10,5))    # pour définir la taille de la figure\n",
    "nx.draw_networkx(graph)\n",
    "plt.axis('off')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### L'agent et le SMA\n",
    "Pour créer un SMA on lui passe simplement le nombre d'agents souhaité. C'est le SMA qui crée les agents. La méthode `set_resources` permet d'affecter les ressources aux agents (via une liste de listes décrivant les `bag` des agents). La méthode `set_accointances` établit la structure du réseau social via une matrice d'adjacence (sous forme d'une liste de listes obtenue facilement à partir d'un graphe networkx).\n",
    "L'agent est initialement créé sans accointance ni ressource, mais il a déjà sa fonction d'utilité.\n",
    "Attention : dans le code suivant, recréer un SMA recrée aussi les fonctions d'utilité de chacun !\n",
    "\n",
    "Un comportement d'agent dans ce cadre consiste à définir 3 aspects :\n",
    "1. Ce qu'il peut mettre en jeu dans un échange (`get_possible_bundles`) ;\n",
    "    Un agent pourrait borner le nombre de ressources qu'il implique lors d'un même échange ou vouloir à tout prix conserver certaines ressources. Il peut donc préparer une liste de propositions contenant ce qu'il est prêt à échanger et l'ordonner pour, par exemple, proposer à l'échanger en premier ce qui a le moins de valeur pour lui. Un `bundle` est sous-ensemble des ressources de son `bag` qu'il est prêt à proposer lors d'un échange. C'est un lot qui peut donc être constitué de plusieurs ressources.\n",
    "\n",
    "1. La manière dont il choisit avec qui il échange et ce qu'il échange (`deal_design`);\n",
    "    Parmi tous ses amis et toutes ses propositions potentielles, un agent pourrait décider de favoriser ses propositions en tentant la première avec tous ses amis avant d'en considérer une autre (`priority_resource`) ou l'inverse, de tenter toutes ses propositions avec un même ami avant d'essayer avec un autre (`priority_friend`), voire de choisir aléatoirement la proposition et l'ami (`random_friend_and_resource`).\n",
    "\n",
    "1. Le critère sur lequel il se base pour décider si ce qu'on lui propose en retour est acceptable (`acceptability`).\n",
    "    Un agent va mettre en avant une de ses propositions et son ami avec lequel il discute aussi. Il doit prendre en compte ce qu'il \"perd\" et ce qu'il gagne pour déterminer si ça vaut le coup selon lui. Il peut se montrer cupide ou égoïste (`selfish`) et toujours vouloir avoir plus (un don n'a pas de sens pour lui), généreux (`social`) et accepter un échange si la société y gagne (même s'il y perd individuellement), ou ne pas faire attention et accepter tout ce qu'on leur propose (`irrational`).\n",
    "\n",
    "\n",
    "Dans l'implémentation qui suit et pour des raisons de simplification, on fait appel à l'objet `sma` pour connaître le bien être social, ce qui est contradictoire avec la décentralisation du processus.\n",
    "Notons que pour le calcul du bien-être social, connaître la variation des utilités des 2 agents qui s'échangent des ressources est parfois suffisant pour en déduire que le bien être social (donc de l'ensemble du système) est amélioré ou pas. Ceci fait l'objet de l'exercice 4."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "%run socialWelfareSrc.py\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Une première expérience\n",
    "\n",
    "Comme première expérience, nous utilisons un graphe social complet (chaque agent connaît toute la société) et nous nous basons sur le comportement le plus simple qui soit :\n",
    "- les agents peuvent mettre en jeu une seule ressource à la fois\n",
    "- ils sont irrationnels (ils acceptent n'importe quelle proposition)\n",
    "- ils choisissent aléatoirement l'article qu'ils échangent et l'ami avec qui ils échangent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "import networkx as nx\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# paramétrage de l'expérience\n",
    "nb_agents=5\n",
    "nb_resources=10     # intervalle de valuation de chaque ressource [1, nb_resources]\n",
    "valeurs = list(range(1,20))\n",
    "# génération du graphe social\n",
    "graph = nx.complete_graph(nb_agents)    \n",
    "m = nx.to_numpy_array(graph, nodelist=range(nb_agents))\n",
    "# génération d'une liste de ressources\n",
    "resources = [genere_nom() for _ in range(nb_resources)]\n",
    "\n",
    "# création du SMA\n",
    "sma = SMA(nb_agents, welfare='Utilitarist', dealDesign='random_friend_and_resource', swap_kind='irrational', resources = resources, valeurs = valeurs, maxEx = 1)\n",
    "sma.set_accointances(m)\n",
    "sma.set_resources(resources)\n",
    "\n",
    "# lancement de l'expérience\n",
    "sma.run(100)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Résultats et interprétation\n",
    "On peut commencer par regarder l'évolution du bien-être social utilitaire. Puisque les agents choisissent aléatoirement et qu'ils acceptent toujours, le bien-être social va fluctuer de manière erratique.\n",
    "\n",
    "On peut également regarder l'évolution du bien-être individuel de chacun des agents. Les agents se comportant de manière irrationnel, les bien-êtres individuels des agents vont également être erratique."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.plot(sma.history,'-o')\n",
    "plt.grid()\n",
    "plt.show()\n",
    "\n",
    "#### Rajouter la seconde courbe sur les 5 utilités dans le temps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#### Une seconde expérience\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sma1 = sma\n",
    "data=pd.DataFrame()\n",
    "for i in range(1) :  ### elle sert à quoi cette boucle ???\n",
    "    # pour tous les comportements possibles\n",
    "    for b in ['random_friend_and_resource','priority_friend','priority_resource'] :\n",
    "        # pour tous les critères d'acceptabilité possibles\n",
    "        for s in ['irrational','social','rational'] :\n",
    "            sma2 = copy.deepcopy(sma1)\n",
    "            sma2.redefineSMA()\n",
    "            sma2.setGlobalMaxEx(i+1)\n",
    "            sma2.setGlobalBehaviorKind(b)\n",
    "            sma2.setGlobalSwapKind(s)\n",
    "            sma2.run(100)\n",
    "            data[b+'_'+s+\"_\"+str(i+1)]=sma2.history\n",
    "            data[b+'_'+s+\"_\"+str(i+1)].plot(figsize=(15,8), grid=True, alpha=1, linestyle = '-')\n",
    "plt.legend(loc=3,bbox_to_anchor=(1,0))\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Tracé de l'évolution du social welfare\n",
    "Une fois l'expérience réalisée, il est alors possible de tracer l'évolution du social Welfare"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On peut bien sûr comparer les comportements en jouant sur les paramètres `dealDesign` et/ou `swap_kind`. Afin de s'assurer que l'on utilise exactement les mêmes affectations d'amis et de ressources, on clone le sma à chaque tour.\n",
    "On prendra garde à la constitution des points de la courbe : si on loggue l'évolution du welfare à chaque changement (dans les méthodes `swap` par exemple), il n'y aura pas les mêmes valeurs collectées; pour assurer le même nombre de valeurs et donc des courbes comparables, on ne loggue qu'à chaque fin de tick (donc dans `runOnce`)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exercices\n",
    "\n",
    "1. Dans le code précédent, le comportement est soit swap soit don. Or le don n'est jamais qu'un swap avec une ressource vide. Comment modifier le swap pour qu'il intègre le don ? Bien évidemment ce don n'a pas de sens pour un agent rationnel puisque si les utilités sont toutes positives, l'agent fait baisser son utilité en donnant.\n",
    "1. Comme le hasard intervient dans chaque expérience (distribution initiale des ressources, graphe s'il n'est pas complet), il est important de pouvoir répéter les mesures afin de moyenner les résultats. \n",
    "Tracer un tableau indiquant le nombre de victoires pour les 6 cas précédents répétés 100 fois. \n",
    "1. Précédemment, nous avons présenté 4 mesures de bien-être social. Que donne le Leximin ? Le leximin consiste à maximiser dans l'ordre lexicographique les vecteurs des utilités de chaque agent classées du min au max\n",
    "1. Dans le code du swap social, on teste l'augmentation de l'utilité sociale en utilisant la classe `sma`. A priori l'agent ne connait que ses accointances. Il ne devrait pas pouvoir accès à tous les agents, donc pas à la classe sma. On triche !. Comment mettre en place des critères locaux permettant de savoir si le bien être social augmente sans avoir accès à tout le monde ?\n",
    "1. le type de graphe\n",
    "    - Est-ce que des phénomènes spécifiques se produisent selon la forme du graphe ou le taux de connectivité moyen ?\n",
    "1. le type d'échange accepté\n",
    "    - Actuellement il n'y a que des échanges de taille 0 ou 1. Qu'est-ce que cela change d'accepter des échanges de 2 ou 3 ressources en même temps ?\n",
    "1. Les agents sont pour l'instant tous homogènes. Ils appliquent tous le même comportement. Qu'est-ce que ça donne si la population est hétérogène ?\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Bibliographie\n",
    "- Moulin, H. (2004). Fair division and collective welfare. MIT press.\n",
    "- Chevaleyre Y, Endriss U, Lang J, Maudet N (2005) Negotiating over small bundles of resources. In: AAMAS’05, pp 296–302\n",
    "- Nongaillard, A., & Mathieu, P. (2014). Agent-based reallocation problem on social networks. Group Decision and Negotiation, 23(5), 1067-1083.\n",
    "- Delahaye, J. P., & Mathieu, P. (2009). La répartition idéale des biens existe-t-elle?. Pour la science, 381, 88-93.\n",
    "- Ferber J. (1995). Systèmes Multi-Agents, vers une intelligence collective. InterEditions. [en ligne ici](http://www.lirmm.fr/~ferber/publications/LesSMA_Ferber.pdf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Leximin Social Welfare\t\t [6, 14, 16, 22, 26]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Leximin Social Welfare\\t\\t', sorted([a.welfare() for a in agentList]) )\n",
    " \n",
    "# 1 if a<b ; -1 if b<a ; 0 if a==b\n",
    "def leximin(a,b) :\n",
    "    for i in range(len(a)) :\n",
    "        if (a[i]>b[i]) :\n",
    "           return 1\n",
    "        elif (a[i]<b[i]) :\n",
    "           return -1\n",
    "    return 0\n",
    "\n",
    "x=np.array([1,2,3,4])\n",
    "y=np.array([1,1,3,4])\n",
    "leximin(x,y) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#****** ANCIENNE VERSION ******\n",
    "\n",
    "class SMA:\n",
    "    def __init__(self, nb_agents, welfare, dealDesign , swap_kind):\n",
    "        self.agentList = [Agent(i, self, dealDesign, swap_kind) for i in range(nb_agents)]\n",
    "        self.welfare_type=welfare\n",
    "        self.tick=0\n",
    "        self.history=[]\n",
    "\n",
    "    def __str__(self) :\n",
    "        s=\"\"\n",
    "        for i in range(len(self.agentList)):\n",
    "            s=s+(self.agentList[i].__str__()+\"\\n\")\n",
    "        s=s+(\"Social Welfare : \" + str(self.socialWelfare())) \n",
    "        return(s)\n",
    "\n",
    "    def set_ressources(self,init_affect) :\n",
    "        if (len(init_affect) != len(self.agentList)) :\n",
    "            raise ValueError(\"Problem of size in set_ressources :\", len(init_affect), len(agentList))\n",
    "        for i in range(len(self.agentList)) :\n",
    "            self.agentList[i].bag=init_affect[i]\n",
    "\n",
    "    def set_accointances(self,adjacency_matrix) :\n",
    "        if (len(adjacency_matrix[0]) != len(self.agentList)) :\n",
    "            raise ValueError(\"Problem of size in set_accointances :\", len(adjacency_matrix[0]) , len(agentList))\n",
    "        for i in range(len(self.agentList)) :\n",
    "            self.agentList[i].accointances = np.nonzero(adjacency_matrix[i])[0]\n",
    "\n",
    "    def run(self, rounds):\n",
    "        self.tick=0\n",
    "        self.history=[]\n",
    "        self.history.append(self.socialWelfare())\n",
    "        for i in range(0,rounds):\n",
    "            self.runOnce()\n",
    "    \n",
    "    def runOnce(self):\n",
    "        self.tick += 1\n",
    "        for agent in self.agentList:\n",
    "            agent.decide(self.tick)\n",
    "        print(\"tick \" + str(self.tick) + \" ended\")\n",
    "        self.history.append(self.socialWelfare())\n",
    "        print(\"Le Welfare actuel est de \", self.socialWelfare())\n",
    "\n",
    "    def socialWelfare(self):\n",
    "        if self.welfare_type.upper()=='UTILITARIST':\n",
    "            return sum([a.welfare() for a in self.agentList])\n",
    "        elif self.welfare_type.upper()=='EGALITARIST':\n",
    "            return min([a.welfare() for a in self.agentList])\n",
    "        elif self.welfare_type.upper()=='ELITIST':\n",
    "            return max([a.welfare() for a in self.agentList])\n",
    "        elif self.welfare_type.upper()=='NASH':\n",
    "            return np.prod([a.welfare() for a in self.agentList])\n",
    "        else :\n",
    "            raise ValueError(\"Unknown method in socialWelfare\")\n",
    "\n",
    "\n",
    "\n",
    "# L'agent est le même que précédemment excepté qu'il a maintenant ses propres accointances, sa fonction d'Utilité\n",
    "# et une méthode de décision avec différents échanges possibles\n",
    "# Par défaut il n'a ni ressources ni accointances\n",
    "class Agent:\n",
    "    def __init__(self, name, sma, dealDesign, swap_kind) :\n",
    "        self.name=name\n",
    "        self.sma = sma\n",
    "        self.swap_kind = swap_kind\n",
    "        self.dealDesign = dealDesign\n",
    "        self.fUtil = {r : random.choice(valeurs) for r in ressources}\n",
    "        self.bag = []\n",
    "        self.accointances = []\n",
    "    def __str__(self) :\n",
    "        return \"agent \"+str(self.name)+\" Welfare :\"+ str(self.welfare())+\"\\tbag :\"+ str(self.bag)\n",
    "    def welfare(self):\n",
    "        return sum([self.fUtil.get(r) for r in self.bag])\n",
    "    def getSortedBag(self):\n",
    "        bagval = [ (r,self.fUtil.get(r)) for r in self.bag]\n",
    "        return [r for r,_ in sorted(bagval, key=lambda x: x[1])]\n",
    "\n",
    "    def gift(self):    \n",
    "        # choisir une accointance au hasard\n",
    "        friend = self.sma.agentList[np.random.choice(self.accointances)]\n",
    "        # choisir une de ses propres ressources au hasard\n",
    "        r = np.random.choice(self.bag)\n",
    "        # donner cette ressource à l'accointance\n",
    "        self.bag.remove(r)\n",
    "        friend.bag.append(r)\n",
    "        print(self.name, \"donne la ressource \",r,\" à l'agent \",friend.name)\n",
    "\n",
    "    # IRRATIONAL : swap sans aucune contrainte\n",
    "    def swap1(self):\n",
    "        # choisir une accointance au hasard\n",
    "        friend = self.sma.agentList[np.random.choice(self.accointances)]\n",
    "        # choisir une de ses propres ressources au hasard\n",
    "        r = np.random.choice(self.bag)\n",
    "        # choisir une des ressources de l'autre au hasard\n",
    "        if len(friend.bag)==0 :\n",
    "            return\n",
    "        s = np.random.choice(friend.bag)\n",
    "        # echanger les ressources\n",
    "        self.bag.remove(r)\n",
    "        friend.bag.remove(s)\n",
    "        friend.bag.append(r)\n",
    "        self.bag.append(s)\n",
    "        print(self.name, \"echange \",r,\" contre \",s, \" avec \",friend.name)\n",
    "    \n",
    "    # PRIORITE RESSOURCE : on choisit une ressource et on teste tous les amis avec\n",
    "    # trois cas possibles : swap_kind = irrational, rational, social\n",
    "    def swap2(self) :\n",
    "        # je propose mes ressources par intérêt décroissant (donc valeur croissante)\n",
    "        for r in self.getSortedBag() :\n",
    "            # à chacun de mes amis\n",
    "            for f in self.accointances :\n",
    "                friend = self.sma.agentList[f]\n",
    "                for s in  friend.bag :\n",
    "                    # on teste si c'est mieux pour les deux\n",
    "                    # UTILISER sma N'EST PAS TRES \"AGENT\" , NEANMOINS DES CRITERES LOCAUX EXISTENT\n",
    "                    actualSocialWelfare = self.sma.socialWelfare()\n",
    "                    actualMyWelfare = self.welfare()\n",
    "                    actualFriendWelfare = friend.welfare()\n",
    "                    self.bag.remove(r)\n",
    "                    friend.bag.remove(s)\n",
    "                    friend.bag.append(r)\n",
    "                    self.bag.append(s)\n",
    "                    # For irrational (equiv to swap1 ; swap1 is a simplified version)\n",
    "                    if self.swap_kind.upper()=='IRRATIONAL' :\n",
    "                        return\n",
    "                    # For rational welfare\n",
    "                    if self.swap_kind.upper()=='RATIONAL' and self.welfare() > actualMyWelfare and friend.welfare() > actualFriendWelfare :\n",
    "                        print(self.name, \"echange \",r,\" contre \",s, \" avec \",friend.name)\n",
    "                        print(self.swap_kind +\" : Welfare passe de \",actualSocialWelfare,\" à \", self.sma.socialWelfare())\n",
    "                        #self.sma.history.append(self.sma.socialWelfare())\n",
    "                        return                       \n",
    "                    # For social welfare\n",
    "                    if self.swap_kind.upper()=='SOCIAL' and self.sma.socialWelfare() > actualSocialWelfare :\n",
    "                        print(self.name, \"echange \",r,\" contre \",s, \" avec \",friend.name)\n",
    "                        print(self.swap_kind + \" : Welfare passe de \",actualSocialWelfare,\" à \", self.sma.socialWelfare())\n",
    "                        #self.sma.history.append(self.sma.socialWelfare())\n",
    "                        return\n",
    "                    # Si les IF ne fonctionnent pas, on remet tout en place\n",
    "                    self.bag.remove(s)\n",
    "                    friend.bag.remove(r)\n",
    "                    friend.bag.append(s)\n",
    "                    self.bag.append(r)\n",
    "    \n",
    "    # PRIORITE FRIEND : on choisit un ami, et on teste toutes les ressources\n",
    "    # trois cas possibles : swap_kind = irrational, rational, social\n",
    "    def swap3(self) :\n",
    "        # je parcours mes amis\n",
    "        for f in self.accointances :\n",
    "            friend = self.sma.agentList[f]\n",
    "            # puis je propose mes ressources par intérêt décroissant (donc valeur croissante)\n",
    "            for r in self.getSortedBag() :\n",
    "                for s in  friend.bag :\n",
    "                    # on teste si c'est mieux pour les deux\n",
    "                    # UTILISER sma N'EST PAS TRES \"AGENT\" , NEANMOINS DES CRITERES LOCAUX EXISTENT\n",
    "                    actualSocialWelfare = self.sma.socialWelfare()\n",
    "                    actualMyWelfare = self.welfare()\n",
    "                    actualFriendWelfare = friend.welfare()\n",
    "                    self.bag.remove(r)\n",
    "                    friend.bag.remove(s)\n",
    "                    friend.bag.append(r)\n",
    "                    self.bag.append(s)\n",
    "                    # For irrational (equiv to swap1 ; swap1 is a simplified version)\n",
    "                    if self.swap_kind.upper()=='IRRATIONAL' :\n",
    "                        return\n",
    "                    # For rational welfare\n",
    "                    if self.swap_kind.upper()=='RATIONAL' and self.welfare() > actualMyWelfare and friend.welfare() > actualFriendWelfare :\n",
    "                        print(self.name, \"echange \",r,\" contre \",s, \" avec \",friend.name)\n",
    "                        print(self.swap_kind +\" : Welfare passe de \",actualSocialWelfare,\" à \", self.sma.socialWelfare())\n",
    "                        #self.sma.history.append(self.sma.socialWelfare())\n",
    "                        return                       \n",
    "                    # For social welfare\n",
    "                    if self.swap_kind.upper()=='SOCIAL' and self.sma.socialWelfare() > actualSocialWelfare :\n",
    "                        print(self.name, \"echange \",r,\" contre \",s, \" avec \",friend.name)\n",
    "                        print(self.swap_kind + \" : Welfare passe de \",actualSocialWelfare,\" à \", self.sma.socialWelfare())\n",
    "                        #self.sma.history.append(self.sma.socialWelfare())\n",
    "                        return\n",
    "                    # Si les IF ne fonctionnent pas, on remet tout en place\n",
    "                    self.bag.remove(s)\n",
    "                    friend.bag.remove(r)\n",
    "                    friend.bag.append(s)\n",
    "                    self.bag.append(r)\n",
    "\n",
    "    def decide(self,tick):\n",
    "        # pour faire qq chose il faut au moins avoir des ressources et des amis\n",
    "        if (len(self.accointances)==0 or len(self.bag)==0) :\n",
    "            return\n",
    "        # s'écrit avec match-case à partir de python 3.10\n",
    "        if   self.dealDesign.upper()=='GIFT'       : self.gift()\n",
    "        elif self.swap_kind.upper()=='IRRATIONAL' : self.swap1()\n",
    "        elif self.dealDesign.upper()=='PRIORITY_RESOURCE'  : self.swap2()\n",
    "        elif self.dealDesign.upper()=='PRIORITY_FRIEND'     : self.swap3()\n",
    "        else : \n",
    "            raise ValueError(\"Impossible behavior in decide\")\n",
    "                "
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "40d3a090f54c6569ab1632332b64b2c03c39dcf918b08424e98f38b5ae0af88f"
  },
  "kernelspec": {
   "display_name": "Python 3.8.5 64-bit ('base': conda)",
   "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": "3.8.10"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}