算法优化

src/main/java/com/aps/entity/Algorithm/ScheduleParams.java

@@ -14,8 +14,8 @@ public class ScheduleParams {
     private LocalDateTime baseTime ; // 当前基准时间
 
     // 基础参数（自适应调整的基准值）
-    private static final int MIN_POPULATION_SIZE = 50;
-    private static final int MAX_POPULATION_SIZE = 100;
+    private static final int MIN_POPULATION_SIZE = 10;
+    private static final int MAX_POPULATION_SIZE = 50;
     private static final int MIN_MAX_ITERATIONS = 50;
     private static final int MAX_MAX_ITERATIONS = 250;
     private static final float MIN_CROSSOVER_PROB = 0.6f;
@@ -54,7 +54,7 @@ public class ScheduleParams {
     private double rsRatio = 0.2; // 随机选择占比
 
     // 自适应调整系数
-    private float populationSizeCoeff = 0.05f; // 种群大小对工序数的敏感系数
+    private float populationSizeCoeff = 0.01f; // 种群大小对工序数的敏感系数
     private float maxIterationsCoeff = 0.08f; // 迭代次数对工序数的敏感系数
     private float crossoverProbCoeff = 0.0001f; // 交叉概率对工序数的敏感系数
     private float mutationProbCoeff = 0.00005f; // 变异概率对工序数的敏感系数

src/main/java/com/aps/service/Algorithm/GeneticDecoder.java

@@ -372,7 +372,7 @@ if(finishedOrder==null||finishedOrder.size()==0)
         int cpuCores = Runtime.getRuntime().availableProcessors();
 //
 //        // 根据工序数量和CPU核心数决定是否使用内部并行
-        if (operationCount > 500 && cpuCores > 4) {
+        if (operationCount > 5000 && cpuCores > 4) {
             // 使用设备并行处理
           //  FileHelper.writeLogFile("使用并行处理 _s");
             parallelDecodeByMachine(chromosome);

src/main/java/com/aps/service/Algorithm/HybridAlgorithm.java

@@ -110,7 +110,7 @@ int opcount=allOperations.size();
         List<GlobalOperationInfo> globalOpList = initialization.generateGlobalOpList();
         // 初始化变邻域搜索
         _vns = new VariableNeighborhoodSearch(_GlobalParam, allOperations, globalOpList, _fitnessCalculator, _objectiveWeights);
-        _hillClimbing = new HillClimbing(_GlobalParam, allOperations, globalOpList, _fitnessCalculator, _objectiveWeights);
+        _hillClimbing = new HillClimbing(allOperations);
         _simulatedAnnealing = new SimulatedAnnealing(_GlobalParam, allOperations, globalOpList, _fitnessCalculator, _objectiveWeights);
         _parallelLocalSearch = new ParallelLocalSearch(_GlobalParam, allOperations, globalOpList, _fitnessCalculator, _objectiveWeights);
         _tabuSearch = new TabuSearch(_GlobalParam, allOperations, globalOpList, _fitnessCalculator, _objectiveWeights);
@@ -146,7 +146,7 @@ int opcount=allOperations.size();
             FileHelper.writeLogFile("爬山法局部优化-----------开始-------");
             GeneticDecoder hillClimbingDecoder = new GeneticDecoder(_GlobalParam, param.getBaseTime(), machines, orders, materials, machineScheduler, materialRequirementService, sceneId);
 
-            Chromosome optimized = _hillClimbing.BatchSearchAll(population, hillClimbingDecoder, param);
+            Chromosome optimized = _hillClimbing.BatchSearchAll(population,machines, hillClimbingDecoder);
             FileHelper.writeLogFile("爬山法局部优化-----------结束-------");
             return getBestChromosome(optimized, param.getBaseTime(), starttime);
         }
@@ -154,14 +154,16 @@ int opcount=allOperations.size();
 
 
         // 步骤2：对初始种群进行模拟退火+爬山法优化
-        FileHelper.writeLogFile("模拟退火+爬山法优化-----------开始-------");
-        GeneticDecoder saDecoder1 = new GeneticDecoder(_GlobalParam, param.getBaseTime(), machines, orders, materials, machineScheduler, materialRequirementService, sceneId);
 
-        List<Chromosome> saHcOptimized = _simulatedAnnealing.batchSearch(population, saDecoder1, param);
+        GeneticDecoder saDecoder1 = new GeneticDecoder(_GlobalParam, param.getBaseTime(), machines, orders, materials, machineScheduler, materialRequirementService, sceneId);
+        if(population.size()>5||opcount<2000 ) {
+            FileHelper.writeLogFile("模拟退火+爬山法优化-----------开始-------");
+            Chromosome saHcOptimized = _simulatedAnnealing.batchSearchGetMax(population, saDecoder1, machines);
+            FileHelper.writeLogFile("模拟退火+爬山法优化-----------结束-------");
 
+            return getBestChromosome(saHcOptimized, param.getBaseTime(), starttime);
+        }
 
-        population = saHcOptimized;
-        FileHelper.writeLogFile("模拟退火+爬山法优化-----------结束-------");
 
         List<List<Chromosome>> combinedFronts = _nsgaIIUtils.parallelFastNonDominatedSort(population);
         int ordercount = globalOpList.stream()
@@ -230,15 +232,6 @@ int opcount=allOperations.size();
             selected.addAll(saImproved);
             FileHelper.writeLogFile("模拟退火搜索-----------结束-------");
 
-            // 爬山法局部优化
-            FileHelper.writeLogFile("爬山法局部优化-----------开始-------");
-            List<Chromosome> hcImproved = new ArrayList<>();
-            for (Chromosome chromosome : selected) {
-                Chromosome hcOptimized = _hillClimbing.search(chromosome, saDecoder, param);
-                hcImproved.add(hcOptimized);
-            }
-            selected.addAll(hcImproved);
-            FileHelper.writeLogFile("爬山法局部优化-----------结束-------");
 
             // 多种局部搜索算法并行
             FileHelper.writeLogFile("并行局部搜索-----------开始-------");
@@ -444,11 +437,6 @@ int opcount=allOperations.size();
             if (existing == null) {
                 // 第一次遇到这个Objectives，直接添加
                 objectivesMap.put(objectivesKey, chromosome);
-            } else {
-                // 已经存在，保留fitness更大的那个
-                if (chromosome.getFitness() > existing.getFitness()) {
-                    objectivesMap.put(objectivesKey, chromosome);
-                }
             }
         }
 
@@ -555,9 +543,9 @@ int opcount=allOperations.size();
         //   _sceneService.saveChromosomeToFile(chromosome, "12345679");
 
         decoder.decodeChromosomeWithCache(chromosome);
-        if (chromosome.getFitness() == 0) {
-            chromosome.setFitness(_fitnessCalculator.calculateFitness(chromosome, _objectiveWeights));
-        }
+//        if (chromosome.getFitness() == 0) {
+//            chromosome.setFitness(_fitnessCalculator.calculateFitness(chromosome, _objectiveWeights));
+//        }
     }
 
 

src/main/java/com/aps/service/Algorithm/ParallelLocalSearch.java

@@ -36,10 +36,10 @@ public class ParallelLocalSearch {
         List<Callable<Chromosome>> tasks = new ArrayList<>();
 
         // 添加爬山法任务
-        tasks.add(() -> {
-            HillClimbing hillClimbing = new HillClimbing(globalParam, allOperations, globalOpList, fitnessCalculator, objectiveWeights);
-            return hillClimbing.search(chromosome, decoder, param);
-        });
+//        tasks.add(() -> {
+//            HillClimbing hillClimbing = new HillClimbing(globalParam, allOperations, globalOpList, fitnessCalculator, objectiveWeights);
+//            return hillClimbing.search(chromosome, decoder, param);
+//        });
 
         // 添加变邻域搜索任务
         tasks.add(() -> {

src/main/java/com/aps/service/Algorithm/SimulatedAnnealing.java

 package com.aps.service.Algorithm;
 
+import com.aps.common.util.FileHelper;
 import com.aps.common.util.ProductionDeepCopyUtil;
 import com.aps.entity.Algorithm.*;
 import com.aps.entity.basic.Entry;
 import com.aps.entity.basic.GlobalParam;
+import com.aps.entity.basic.Machine;
 
 import java.util.*;
 import java.util.concurrent.*;
@@ -33,39 +35,59 @@ public class SimulatedAnnealing {
             new ArrayBlockingQueue<>(200), // 有界队列，避免内存溢出
             new ThreadPoolExecutor.CallerRunsPolicy() // 任务满了主线程执行，不丢失任务
     );
-    public List<Chromosome> batchSearch(List<Chromosome> chromosomes, GeneticDecoder decoder, ScheduleParams param) {
+    public List<Chromosome> batchSearch(List<Chromosome> chromosomes, GeneticDecoder decoder, List<Machine> machines) {
         List<Chromosome> saHcOptimized=new ArrayList<>();
-        CompletableFuture.allOf(chromosomes.stream()
-                        .map(chromosome -> CompletableFuture.runAsync(() -> {
-                            Chromosome optimized = searchWithHillClimbing(chromosome, decoder, param);
+//        CompletableFuture.allOf(chromosomes.stream()
+//                        .map(chromosome -> CompletableFuture.runAsync(() -> {
+//                            Chromosome optimized = searchWithHillClimbing(chromosome, decoder, param);
+//                            saHcOptimized.add(optimized);
+//                        }, decodeExecutor))
+//                        .toArray(CompletableFuture[]::new))
+//                .join();
+        for (Chromosome chromosome:chromosomes) {
+            Chromosome optimized = searchWithHillClimbing(chromosome, decoder, machines);
                             saHcOptimized.add(optimized);
-                        }, decodeExecutor))
-                        .toArray(CompletableFuture[]::new))
-                .join();
+                            break;
+        }
         return saHcOptimized;
     }
 
+    public Chromosome batchSearchGetMax(List<Chromosome> chromosomes, GeneticDecoder decoder, List<Machine> machines) {
+        List<Chromosome> saHcOptimized=batchSearch(chromosomes,decoder,machines);
+
+        int bestidx= Getbest(saHcOptimized,null);
+        if(bestidx>-1) {
+            return saHcOptimized.get(bestidx);
+        }
+        return  null;
+
+    }
+
+
         /**
          * 模拟退火搜索，当温度降低到一定程度后切换到爬山法
          * 流程：模拟退火迭代 → 随机生成新排产 → 按概率接受 → 降温 → 温度低时爬山法求精 → 输出最优
          */
-    public Chromosome searchWithHillClimbing(Chromosome chromosome, GeneticDecoder decoder, ScheduleParams param) {
+    public Chromosome searchWithHillClimbing(Chromosome chromosome, GeneticDecoder decoder, List<Machine> machines) {
         Chromosome current = ProductionDeepCopyUtil.deepCopy(chromosome, Chromosome.class);
         Chromosome best = ProductionDeepCopyUtil.deepCopy(chromosome, Chromosome.class);
 
-        // 初始化解码
-     //   decode(decoder, best, param);
-     //   decode(decoder, current, param);
+
         Map<String, Entry> Entrys = buildEntryKey(current);
+        FileHelper.writeLogFile("模拟退火+爬山法 - 初始化索引完成");
 
         // 初始化温度 - 优化速度：更快降温，更快切换到爬山法
         double temperature = 100.0;
         double coolingRate = 0.95;
         double temperatureThreshold = 1.0; // 提高阈值，更快切换到爬山法
         int maxIterations = 300; // 减少最大迭代次数
+        FileHelper.writeLogFile(String.format("模拟退火+爬山法 - 参数配置：温度=%.1f, 降温率=%.2f, 阈值=%.1f, 最大迭代=%d",
+                temperature, coolingRate, temperatureThreshold, maxIterations));
+
         int noImproveCount = 0;
         int maxNoImprove = 50; // 连续50次没改进就提前停止
-
+        int acceptCount = 0;
+        int improveCount = 0;
         for (int i = 0; i < maxIterations; i++) {
             boolean improved = false;
 
@@ -73,7 +95,7 @@ public class SimulatedAnnealing {
             Chromosome neighbor = generateNeighbor(current,Entrys);
 
             // 2. 解码并计算适应度（只解码neighbor，current已经解码过了）
-            decode(decoder, neighbor, param);
+            decode(decoder, neighbor, machines);
 
             // 3. 计算能量差（fitness差）
             double energyDifference = calculateEnergyDifference(neighbor, current);
@@ -81,13 +103,16 @@ public class SimulatedAnnealing {
             // 4. 按概率接受新解
             // - 更优：直接接受
             // - 较差：概率接受（高温更容易接受）
+            boolean accepted = false;
             if (energyDifference > 0 || rnd.nextDouble() < Math.exp(energyDifference / temperature)) {
                 current = neighbor;
-
+                accepted = true;
+                acceptCount++;
                 // 更新全局最优
                 if (isBetter(current, best)) {
                     best = ProductionDeepCopyUtil.deepCopy(current, Chromosome.class);
                     improved = true;
+                    improveCount++;
                     noImproveCount = 0;
                 }
             }
@@ -98,16 +123,27 @@ public class SimulatedAnnealing {
 
             // 5. 降温
             temperature *= coolingRate;
-
+           if ((i + 1) % 10 == 0) {
+                FileHelper.writeLogFile(String.format("模拟退火+爬山法 - 迭代%d/%d：温度=%.4f, 接受数=%d, 改进数=%d, 无改进连续=%d",
+                        i + 1, maxIterations, temperature, acceptCount, improveCount, noImproveCount));
+            }
             // 6. 提前停止条件：温度低于阈值 或 连续多次没改进
             if (temperature < temperatureThreshold || noImproveCount >= maxNoImprove) {
-                HillClimbing hillClimbing = new HillClimbing(globalParam, allOperations, globalOpList, fitnessCalculator, objectiveWeights);
-                Chromosome refined = hillClimbing.search(best, decoder, param);
+                String stopReason = temperature < temperatureThreshold ? "温度低于阈值" : "连续无改进达到上限";
+                FileHelper.writeLogFile(String.format("模拟退火+爬山法 - 提前停止：%s，迭代%d次，最终温度=%.4f",
+                        stopReason, i + 1, temperature));
+
+                FileHelper.writeLogFile("模拟退火+爬山法 - 切换到爬山法求精");
+
+                HillClimbing hillClimbing = new HillClimbing( allOperations );
+                Chromosome refined = hillClimbing.search(best, decoder,machines);
+                FileHelper.writeLogFile("模拟退火+爬山法 - 爬山法求精完成");
 
                 // 返回爬山法求精后的结果
                 return refined;
             }
         }
+        FileHelper.writeLogFile(String.format("模拟退火+爬山法 - 完成所有%d次迭代，最终fitness=%.4f", maxIterations, best.getFitness()));
 
         // 7. 输出全局最优排产
         return best;
@@ -321,11 +357,15 @@ public class SimulatedAnnealing {
     /**
      * 解码染色体
      */
-    private void decode(GeneticDecoder decoder, Chromosome chromosome, ScheduleParams param) {
+    private void decode(GeneticDecoder decoder, Chromosome chromosome , List<Machine> machines) {
+        chromosome.setResult(new CopyOnWriteArrayList<>());
+
+        // 假设Machine类有拷贝方法，或使用MapStruct等工具进行映射
+
+        chromosome.setMachines(ProductionDeepCopyUtil.deepCopyList(machines, Machine.class)); // 简单拷贝，实际可能需要深拷贝
+
         decoder.decodeChromosomeWithCache(chromosome);
-        if (chromosome.getFitness() == 0) {
-            chromosome.setFitness(fitnessCalculator.calculateFitness(chromosome, objectiveWeights));
-        }
+
     }
 
     /**
@@ -341,4 +381,23 @@ public class SimulatedAnnealing {
         }
         return false;
     }
+    private int  Getbest(List<Chromosome> candidates,Chromosome best) {
+        // 找出最佳候选方案
+
+        int bestidx = -1;
+        if(best==null)
+        {
+            best=candidates.get(0);
+            bestidx=0;
+        }
+        for (int i = 0; i < candidates.size(); i++) {
+            Chromosome candidate = candidates.get(i);
+            if (isBetter(candidate, best)) {
+                bestidx = i;
+            }
+        }
+
+        return bestidx;
+    }
+
 }
\ No newline at end of file