|
| 1 | +package DataMining_ACO; |
| 2 | + |
| 3 | +import java.io.BufferedReader; |
| 4 | +import java.io.File; |
| 5 | +import java.io.FileReader; |
| 6 | +import java.io.IOException; |
| 7 | +import java.text.MessageFormat; |
| 8 | +import java.util.ArrayList; |
| 9 | +import java.util.Collections; |
| 10 | +import java.util.HashMap; |
| 11 | +import java.util.Map; |
| 12 | +import java.util.Random; |
| 13 | + |
| 14 | +/** |
| 15 | + * 蚁群算法工具类 |
| 16 | + * |
| 17 | + * @author lyq |
| 18 | + * |
| 19 | + */ |
| 20 | +public class ACOTool { |
| 21 | + // 输入数据类型 |
| 22 | + public static final int INPUT_CITY_NAME = 1; |
| 23 | + public static final int INPUT_CITY_DIS = 2; |
| 24 | + |
| 25 | + // 城市间距离邻接矩阵 |
| 26 | + public static double[][] disMatrix; |
| 27 | + // 当前时间 |
| 28 | + public static int currentTime; |
| 29 | + |
| 30 | + // 测试数据地址 |
| 31 | + private String filePath; |
| 32 | + // 蚂蚁数量 |
| 33 | + private int antNum; |
| 34 | + // 控制参数 |
| 35 | + private double alpha; |
| 36 | + private double beita; |
| 37 | + private double p; |
| 38 | + private double Q; |
| 39 | + // 随机数产生器 |
| 40 | + private Random random; |
| 41 | + // 城市名称集合,这里为了方便，将城市用数字表示 |
| 42 | + private ArrayList<String> totalCitys; |
| 43 | + // 所有的蚂蚁集合 |
| 44 | + private ArrayList<Ant> totalAnts; |
| 45 | + // 城市间的信息素浓度矩阵，随着时间的增多而减少 |
| 46 | + private double[][] pheromoneMatrix; |
| 47 | + // 目标的最短路径,顺序为从集合的前部往后挪动 |
| 48 | + private ArrayList<String> bestPath; |
| 49 | + // 信息素矩阵存储图,key采用的格式(i,j,t)->value |
| 50 | + private Map<String, Double> pheromoneTimeMap; |
| 51 | + |
| 52 | + public ACOTool(String filePath, int antNum, double alpha, double beita, |
| 53 | + double p, double Q) { |
| 54 | + this.filePath = filePath; |
| 55 | + this.antNum = antNum; |
| 56 | + this.alpha = alpha; |
| 57 | + this.beita = beita; |
| 58 | + this.p = p; |
| 59 | + this.Q = Q; |
| 60 | + this.currentTime = 0; |
| 61 | + |
| 62 | + readDataFile(); |
| 63 | + } |
| 64 | + |
| 65 | + /** |
| 66 | + * 从文件中读取数据 |
| 67 | + */ |
| 68 | + private void readDataFile() { |
| 69 | + File file = new File(filePath); |
| 70 | + ArrayList<String[]> dataArray = new ArrayList<String[]>(); |
| 71 | + |
| 72 | + try { |
| 73 | + BufferedReader in = new BufferedReader(new FileReader(file)); |
| 74 | + String str; |
| 75 | + String[] tempArray; |
| 76 | + while ((str = in.readLine()) != null) { |
| 77 | + tempArray = str.split(" "); |
| 78 | + dataArray.add(tempArray); |
| 79 | + } |
| 80 | + in.close(); |
| 81 | + } catch (IOException e) { |
| 82 | + e.getStackTrace(); |
| 83 | + } |
| 84 | + |
| 85 | + int flag = -1; |
| 86 | + int src = 0; |
| 87 | + int des = 0; |
| 88 | + int size = 0; |
| 89 | + // 进行城市名称种数的统计 |
| 90 | + this.totalCitys = new ArrayList<>(); |
| 91 | + for (String[] array : dataArray) { |
| 92 | + if (array[0].equals("#") && totalCitys.size() == 0) { |
| 93 | + flag = INPUT_CITY_NAME; |
| 94 | + |
| 95 | + continue; |
| 96 | + } else if (array[0].equals("#") && totalCitys.size() > 0) { |
| 97 | + size = totalCitys.size(); |
| 98 | + // 初始化距离矩阵 |
| 99 | + this.disMatrix = new double[size + 1][size + 1]; |
| 100 | + this.pheromoneMatrix = new double[size + 1][size + 1]; |
| 101 | + |
| 102 | + // 初始值-1代表此对应位置无值 |
| 103 | + for (int i = 0; i < size; i++) { |
| 104 | + for (int j = 0; j < size; j++) { |
| 105 | + this.disMatrix[i][j] = -1; |
| 106 | + this.pheromoneMatrix[i][j] = -1; |
| 107 | + } |
| 108 | + } |
| 109 | + |
| 110 | + flag = INPUT_CITY_DIS; |
| 111 | + continue; |
| 112 | + } |
| 113 | + |
| 114 | + if (flag == INPUT_CITY_NAME) { |
| 115 | + this.totalCitys.add(array[0]); |
| 116 | + } else { |
| 117 | + src = Integer.parseInt(array[0]); |
| 118 | + des = Integer.parseInt(array[1]); |
| 119 | + |
| 120 | + this.disMatrix[src][des] = Double.parseDouble(array[2]); |
| 121 | + this.disMatrix[des][src] = Double.parseDouble(array[2]); |
| 122 | + } |
| 123 | + } |
| 124 | + } |
| 125 | + |
| 126 | + /** |
| 127 | + * 计算从蚂蚁城市i到j的概率 |
| 128 | + * |
| 129 | + * @param cityI |
| 130 | + * 城市I |
| 131 | + * @param cityJ |
| 132 | + * 城市J |
| 133 | + * @param currentTime |
| 134 | + * 当前时间 |
| 135 | + * @return |
| 136 | + */ |
| 137 | + private double calIToJProbably(String cityI, String cityJ, int currentTime) { |
| 138 | + double pro = 0; |
| 139 | + double n = 0; |
| 140 | + double pheromone; |
| 141 | + int i; |
| 142 | + int j; |
| 143 | + |
| 144 | + i = Integer.parseInt(cityI); |
| 145 | + j = Integer.parseInt(cityJ); |
| 146 | + |
| 147 | + pheromone = getPheromone(currentTime, cityI, cityJ); |
| 148 | + n = 1.0 / disMatrix[i][j]; |
| 149 | + |
| 150 | + if (pheromone == 0) { |
| 151 | + pheromone = 1; |
| 152 | + } |
| 153 | + |
| 154 | + pro = Math.pow(n, alpha) * Math.pow(pheromone, beita); |
| 155 | + |
| 156 | + return pro; |
| 157 | + } |
| 158 | + |
| 159 | + /** |
| 160 | + * 计算综合概率蚂蚁从I城市走到J城市的概率 |
| 161 | + * |
| 162 | + * @return |
| 163 | + */ |
| 164 | + public String selectAntNextCity(Ant ant, int currentTime) { |
| 165 | + double randomNum; |
| 166 | + double tempPro; |
| 167 | + // 总概率指数 |
| 168 | + double proTotal; |
| 169 | + String nextCity = null; |
| 170 | + ArrayList<String> allowedCitys; |
| 171 | + // 各城市概率集 |
| 172 | + double[] proArray; |
| 173 | + |
| 174 | + // 如果是刚刚开始的时候，没有路过任何城市，则随机返回一个城市 |
| 175 | + if (ant.currentPath.size() == 0) { |
| 176 | + nextCity = String.valueOf(random.nextInt(totalCitys.size()) + 1); |
| 177 | + |
| 178 | + return nextCity; |
| 179 | + } else if (ant.nonVisitedCitys.isEmpty()) { |
| 180 | + // 如果全部遍历完毕，则再次回到起点 |
| 181 | + nextCity = ant.currentPath.get(0); |
| 182 | + |
| 183 | + return nextCity; |
| 184 | + } |
| 185 | + |
| 186 | + proTotal = 0; |
| 187 | + allowedCitys = ant.nonVisitedCitys; |
| 188 | + proArray = new double[allowedCitys.size()]; |
| 189 | + |
| 190 | + for (int i = 0; i < allowedCitys.size(); i++) { |
| 191 | + nextCity = allowedCitys.get(i); |
| 192 | + proArray[i] = calIToJProbably(ant.currentPos, nextCity, currentTime); |
| 193 | + proTotal += proArray[i]; |
| 194 | + } |
| 195 | + |
| 196 | + for (int i = 0; i < allowedCitys.size(); i++) { |
| 197 | + // 归一化处理 |
| 198 | + proArray[i] /= proTotal; |
| 199 | + } |
| 200 | + |
| 201 | + // 用随机数选择下一个城市 |
| 202 | + randomNum = random.nextInt(100) + 1; |
| 203 | + randomNum = randomNum / 100; |
| 204 | + // 因为1.0是无法判断到的，,总和会无限接近1.0取为0.99做判断 |
| 205 | + if (randomNum == 1) { |
| 206 | + randomNum = randomNum - 0.01; |
| 207 | + } |
| 208 | + |
| 209 | + tempPro = 0; |
| 210 | + // 确定区间 |
| 211 | + for (int j = 0; j < allowedCitys.size(); j++) { |
| 212 | + if (randomNum > tempPro && randomNum <= tempPro + proArray[j]) { |
| 213 | + // 采用拷贝的方式避免引用重复 |
| 214 | + nextCity = allowedCitys.get(j); |
| 215 | + break; |
| 216 | + } else { |
| 217 | + tempPro += proArray[j]; |
| 218 | + } |
| 219 | + } |
| 220 | + |
| 221 | + return nextCity; |
| 222 | + } |
| 223 | + |
| 224 | + /** |
| 225 | + * 获取给定时间点上从城市i到城市j的信息素浓度 |
| 226 | + * |
| 227 | + * @param t |
| 228 | + * @param cityI |
| 229 | + * @param cityJ |
| 230 | + * @return |
| 231 | + */ |
| 232 | + private double getPheromone(int t, String cityI, String cityJ) { |
| 233 | + double pheromone = 0; |
| 234 | + String key; |
| 235 | + |
| 236 | + // 上一周期需将时间倒回一周期 |
| 237 | + key = MessageFormat.format("{0},{1},{2}", cityI, cityJ, t); |
| 238 | + |
| 239 | + if (pheromoneTimeMap.containsKey(key)) { |
| 240 | + pheromone = pheromoneTimeMap.get(key); |
| 241 | + } |
| 242 | + |
| 243 | + return pheromone; |
| 244 | + } |
| 245 | + |
| 246 | + /** |
| 247 | + * 每轮结束，刷新信息素浓度矩阵 |
| 248 | + * |
| 249 | + * @param t |
| 250 | + */ |
| 251 | + private void refreshPheromone(int t) { |
| 252 | + double pheromone = 0; |
| 253 | + // 上一轮周期结束后的信息素浓度，丛信息素浓度图中查找 |
| 254 | + double lastTimeP = 0; |
| 255 | + // 本轮信息素浓度增加量 |
| 256 | + double addPheromone; |
| 257 | + String key; |
| 258 | + |
| 259 | + for (String i : totalCitys) { |
| 260 | + for (String j : totalCitys) { |
| 261 | + if (!i.equals(j)) { |
| 262 | + // 上一周期需将时间倒回一周期 |
| 263 | + key = MessageFormat.format("{0},{1},{2}", i, j, t - 1); |
| 264 | + |
| 265 | + if (pheromoneTimeMap.containsKey(key)) { |
| 266 | + lastTimeP = pheromoneTimeMap.get(key); |
| 267 | + } else { |
| 268 | + lastTimeP = 0; |
| 269 | + } |
| 270 | + |
| 271 | + addPheromone = 0; |
| 272 | + for (Ant ant : totalAnts) { |
| 273 | + // 每只蚂蚁传播的信息素为控制因子除以距离总成本 |
| 274 | + addPheromone += Q / ant.calSumDistance(); |
| 275 | + } |
| 276 | + |
| 277 | + // 将上次的结果值加上递增的量，并存入图中 |
| 278 | + pheromone = p * lastTimeP + addPheromone; |
| 279 | + key = MessageFormat.format("{0},{1},{2}", i, j, t); |
| 280 | + pheromoneTimeMap.put(key, pheromone); |
| 281 | + } |
| 282 | + } |
| 283 | + } |
| 284 | + |
| 285 | + } |
| 286 | + |
| 287 | + public void antStartSearching() { |
| 288 | + // 蚁群寻找的总次数 |
| 289 | + int loopCount = 0; |
| 290 | + // 选中的下一个城市 |
| 291 | + String selectedCity = ""; |
| 292 | + |
| 293 | + pheromoneTimeMap = new HashMap<String, Double>(); |
| 294 | + totalAnts = new ArrayList<>(); |
| 295 | + random = new Random(); |
| 296 | + |
| 297 | + while (loopCount < 10) { |
| 298 | + initAnts(); |
| 299 | + |
| 300 | + while (true) { |
| 301 | + for (Ant ant : totalAnts) { |
| 302 | + selectedCity = selectAntNextCity(ant, currentTime); |
| 303 | + ant.goToNextCity(selectedCity); |
| 304 | + } |
| 305 | + |
| 306 | + // 如果已经遍历完所有城市，则跳出此轮循环 |
| 307 | + if (totalAnts.get(0).isBack()) { |
| 308 | + break; |
| 309 | + } |
| 310 | + } |
| 311 | + |
| 312 | + // 周期时间叠加 |
| 313 | + currentTime++; |
| 314 | + refreshPheromone(currentTime); |
| 315 | + } |
| 316 | + |
| 317 | + // 根据距离成本，选出所花距离最短的一个路径 |
| 318 | + Collections.sort(totalAnts); |
| 319 | + bestPath = totalAnts.get(0).currentPath; |
| 320 | + for (String cityName : bestPath) { |
| 321 | + System.out.println(MessageFormat.format("-->{0}", cityName)); |
| 322 | + } |
| 323 | + } |
| 324 | + |
| 325 | + /** |
| 326 | + * 初始化蚁群操作 |
| 327 | + */ |
| 328 | + private void initAnts() { |
| 329 | + Ant tempAnt; |
| 330 | + ArrayList<String> nonVisitedCitys; |
| 331 | + totalAnts.clear(); |
| 332 | + |
| 333 | + // 初始化蚁群 |
| 334 | + for (int i = 0; i < antNum; i++) { |
| 335 | + nonVisitedCitys = (ArrayList<String>) totalCitys.clone(); |
| 336 | + tempAnt = new Ant(pheromoneMatrix, nonVisitedCitys); |
| 337 | + |
| 338 | + totalAnts.add(tempAnt); |
| 339 | + } |
| 340 | + } |
| 341 | +} |
0 commit comments