Changeset 11806 for branches/HeuristicLab.Problems.GrammaticalOptimization/HeuristicLab.Algorithms.Bandits/BanditPolicies/GenericThompsonSamplingPolicy.cs

Timestamp:

01/20/15 20:25:00 (9 years ago)

Author:

gkronber

Message:

#2283: separated value-states from done-states in GenericGrammarPolicy and removed disabling of actions from bandit policies

File:

• branches/HeuristicLab.Problems.GrammaticalOptimization/HeuristicLab.Algorithms.Bandits/BanditPolicies/GenericThompsonSamplingPolicy.cs (modified) (1 diff)

branches/HeuristicLab.Problems.GrammaticalOptimization/HeuristicLab.Algorithms.Bandits/BanditPolicies/GenericThompsonSamplingPolicy.cs

@@ -22 +22 @@
       foreach (var aInfo in myActionInfos) {
         aIdx++;
-        if (aInfo.Disabled) continue;
-        //if (aInfo.Tries == 0) return aIdx;
         var q = aInfo.SampleExpectedReward(random);
         if (q > bestQ) {