Changeset 12876 for branches/HeuristicLab.Problems.GrammaticalOptimization-gkr/HeuristicLab.Algorithms.Bandits/ActionInfos

Timestamp:

08/17/15 19:13:19 (9 years ago)

Author:

gkronber

Message:

#2283: implemented first crude version of extreme hunter algorithm in branch

Location:

branches/HeuristicLab.Problems.GrammaticalOptimization-gkr/HeuristicLab.Algorithms.Bandits/ActionInfos

Files:

• DefaultPolicyActionInfo.cs (modified) (1 diff)
• ExtremeHunterActionInfo.cs (added)

branches/HeuristicLab.Problems.GrammaticalOptimization-gkr/HeuristicLab.Algorithms.Bandits/ActionInfos/DefaultPolicyActionInfo.cs

@@ -27 +27 @@
       MaxReward = Math.Max(MaxReward, reward);
       var delta = reward - avgValue;
-      //var alpha = 0.01;
-      var alpha = Math.Max(1.0/Tries, 0.01);
+      double alpha = 1.0 / Tries;
       avgValue = avgValue + alpha * delta;
     }