Pattern_Appro_Weak.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aa5d1b6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from gurobipy import *  #引用gurobi求解包\n",
    "import numpy as np\n",
    "\n",
    "\n",
    "def Pattern_Appro_solver(s,timebound):\n",
    "#Model\n",
    "    OP = Model(\"PA\")\n",
    "#Number of variables\n",
    "    P = len(s.Pi)\n",
    "    J = int(s.number_of_batch)\n",
    "    B = len(s.mb)\n",
    "    numberperunit = int(s.number_of_order / s.number_of_batch)\n",
    "#Add Vars\n",
    "    y = OP.addVars(P, J, vtype=GRB.INTEGER, name=\"y\") \n",
    "    delta = OP.addVars(J, B, vtype=GRB.BINARY, name=\"delta\")\n",
    "    for j in range(J):\n",
    "        for p in range(P):\n",
    "            if s.Qp[p] == 1:\n",
    "                y[p, j].vtype = GRB.BINARY\n",
    "            else:\n",
    "                y[p, j].ub = s.Qp[p]\n",
    "#InitialIndex\n",
    "    for i in s.InitialIndex:\n",
    "        for j, k, l in s.Pi:\n",
    "            if i in k and len(k) == 1:\n",
    "                y[j, s.InitialIndex.index(i)].lb = 1\n",
    "                for t in range(s.InitialIndex.index(i)) or range(\n",
    "                        s.InitialIndex.index(i) + 1, J):\n",
    "                    y[j, t].ub = 0\n",
    "                break\n",
    "            elif i in k:\n",
    "                y[j, s.InitialIndex.index(i)].lb = 1\n",
    "                y[j, s.InitialIndex.index(i)].ub = 1\n",
    "                break\n",
    "\n",
    "\n",
    "#Set objective\n",
    "    OP.setObjective(quicksum(delta[j, b] for j in range(J) for b in range(B)),\n",
    "                    GRB.MINIMIZE)\n",
    "#Add Constr\n",
    "    OP.addConstrs(sum(y[p, j] for p in range(P)) == numberperunit for j in range(J))\n",
    "    OP.addConstrs(sum(y[p, j] for j in range(J)) == s.Qp[p] for p in range(P))\n",
    "    OP.addConstrs(numberperunit * delta[j, b] >= sum(y[p, j] for p in s.Pb[b])\n",
    "                  for j in range(J) for b in range(B))\n",
    "#Set time and Log file\n",
    "    OP.Params.timelimit = timebound\n",
    "    OP.Params.LogToConsole = 0\n",
    "    FileStr = './Log/' + 'Pattern_Appro_solver' + '_' + str(s.number_of_order) + '_' + 'order' + '_' + str(s.number_of_batch) + '_' + 'batch' + '.log'\n",
    "    OP.Params.LogFile = FileStr\n",
    "    \n",
    "    OP.update()\n",
    "    OP.optimize()\n",
    "\n",
    "#Save solution\n",
    "    Pi=copy.deepcopy(s.Pi)\n",
    "    #Pi=s.Pi\n",
    "    OP.setParam(\"SolutionNumber\", 0)\n",
    "    vp = OP.getAttr('xn', y)\n",
    "    vy = np.zeros((s.number_of_order, J))\n",
    "    for p in range(P):\n",
    "        for j in range(J):\n",
    "            if vp[p, j] > 0:\n",
    "                for k in range(round(vp[p, j])):\n",
    "                    vy[Pi[p][1][-1]][j] = 1\n",
    "                    Pi[p][1].pop()\n",
    "    matrix = [[] for j in range(J)]\n",
    "    for j in range(J):\n",
    "        matrix[j].extend(list(np.where(vy[:, j] == 1)[0]))\n",
    "    s.Pattern_Appro_matrix = matrix\n",
    "    return OP.ObjVal,OP.MIPGap"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}