Changeset 11770 for branches/HeuristicLab.Problems.GrammaticalOptimization/Main

Timestamp:

01/15/15 18:59:07 (10 years ago)

Author:

gkronber

Message:

#2283: worked on generic sequential search alg with bandit policy as parameter

File:

• branches/HeuristicLab.Problems.GrammaticalOptimization/Main/Program.cs (modified) (7 diffs)

branches/HeuristicLab.Problems.GrammaticalOptimization/Main/Program.cs

@@ -137 +137 @@
 
     private static void RunDemo() {
+      // TODO: move problem instances into a separate folder
+      // TODO: improve performance of SequentialSearch (memory allocations related to sequences)
+      // TODO: implement bridge to HL-GP 
       // TODO: unify MCTS, TD and ContextMCTS Solvers (stateInfos)
       // TODO: test with eps-greedy using max instead of average as value (seems to work well for symb-reg! explore further!)
@@ -161 +164 @@
       int iterations = 0;
       var sw = new Stopwatch();
-      double bestQuality = 0;
-      string bestSentence = "";
+
       var globalStatistics = new SentenceSetStatistics();
       var random = new Random();
@@ -168 +170 @@
       //var phraseLen = 3;
       //var numPhrases = 5;
-      //var problem = new RoyalPhraseSequenceProblem(random, 10, numPhrases, phraseLen: phraseLen, numCorrectPhrases: 1, correctReward: 1, incorrectReward: 0.0, phrasesAsSets: true);
-
-      //var phraseLen = 4;
-      //var numPhrases = 5;
-      //var problem = new FindPhrasesProblem(random, 15, numPhrases, phraseLen, numOptimalPhrases: numPhrases, numDecoyPhrases: 500, correctReward: 1.0, decoyReward: 0.2, phrasesAsSets: true);
-
-      var problem = new SymbolicRegressionPoly10Problem();   // good results e.g. 10 randomtries and EpsGreedyPolicy(0.2, (aInfo)=>aInfo.MaxReward)
+      //var problem = new RoyalPhraseSequenceProblem(random, 15, numPhrases, phraseLen: phraseLen, numCorrectPhrases: 1, correctReward: 1, incorrectReward: 0.0, phrasesAsSets: true);
+
+      // var phraseLen = 2;
+      // var numPhrases = 5;
+      // var problem = new FindPhrasesProblem(random, 15, numPhrases, phraseLen, numOptimalPhrases: numPhrases, numDecoyPhrases: 0, correctReward: 1.0, decoyReward: 0.0, phrasesAsSets: true);
+
+      //var problem = new SymbolicRegressionPoly10Problem();   // good results e.g. 10 randomtries and EpsGreedyPolicy(0.2, (aInfo)=>aInfo.MaxReward)
       // Ant
       // good results e.g. with       var alg = new MctsSampler(problem, 17, random, 1, (rand, numActions) => new ThresholdAscentPolicy(numActions, 500, 0.01));
       // GaussianModelWithUnknownVariance (and Q= 0.99-quantil) also works well for Ant
-      //var problem = new SantaFeAntProblem();
+      // very good results with:       var alg = new SequentialSearch(problem, 17, random, 0, 
+      // new HeuristicLab.Algorithms.Bandits.GrammarPolicies.GenericGrammarPolicy(problem, new UCB1TunedPolicy(), true));
+
+      var problem = new SantaFeAntProblem();
       //var problem = new SymbolicRegressionProblem("Tower"); 
       //var problem = new PalindromeProblem();
@@ -186 +191 @@
       // symbreg length = 11 q = 0.824522210419616
       //var alg = new MctsSampler(problem, 23, random, 0, new BoltzmannExplorationPolicy(100));
-      var alg = new MctsSampler(problem, 23, random, 0, new EpsGreedyPolicy(0.1));
+      //var alg = new MctsSampler(problem, 23, random, 0, new EpsGreedyPolicy(0.1));
+      var alg = new SequentialSearch(problem, 10, random, 0, 
+        new HeuristicLab.Algorithms.Bandits.GrammarPolicies.GenericGrammarPolicy(problem, new GaussianThompsonSamplingPolicy(true), true));
       //var alg = new MctsQLearningSampler(problem, sentenceLen, random, 0, null);
       //var alg = new MctsQLearningSampler(problem, 30, random, 0, new EpsGreedyPolicy(0.2));
@@ -199 +206 @@
 
       alg.FoundNewBestSolution += (sentence, quality) => {
-        bestQuality = quality;
-        bestSentence = sentence;
         //Console.WriteLine("{0,4} {1,7} {2}", alg.treeDepth, alg.treeSize, globalStatistics);
         //Console.ReadLine();
@@ -208 +213 @@
         globalStatistics.AddSentence(sentence, quality);
         if (iterations % 100 == 0) {
-          //if (iterations % 1000 == 0) Console.Clear();
+          if (iterations % 1000 == 0) Console.Clear();
           Console.SetCursorPosition(0, 0);
           alg.PrintStats();
@@ -226 +231 @@
       sw.Stop();
 
-      Console.WriteLine("{0,10} Best soultion: {1,10:F5} {2}", iterations, bestQuality, bestSentence);
+      Console.Clear();
+      alg.PrintStats();
+      Console.WriteLine(globalStatistics);
       Console.WriteLine("{0:F2} sec {1,10:F1} sols/sec {2,10:F1} ns/sol",
         sw.Elapsed.TotalSeconds,