Changeset 11742 for branches/HeuristicLab.Problems.GrammaticalOptimization/HeuristicLab.Algorithms.Bandits/BanditPolicies/DefaultPolicyActionInfo.cs

Timestamp:

01/09/15 14:57:28 (10 years ago)

Author:

gkronber

Message:

#2283 refactoring

Location:

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

Files:

• . (moved) (moved from branches/HeuristicLab.Problems.GrammaticalOptimization/HeuristicLab.Algorithms.Bandits/Policies)
• DefaultPolicyActionInfo.cs (modified) (2 diffs)

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

@@ -6 +6 @@
 using System.Threading.Tasks;
 
-namespace HeuristicLab.Algorithms.Bandits {
+namespace HeuristicLab.Algorithms.Bandits.BanditPolicies {
   // stores information that is relevant for most of the policies
-  public class DefaultPolicyActionInfo : IPolicyActionInfo {
+  public class DefaultPolicyActionInfo : IBanditPolicyActionInfo {
     public bool Disabled { get { return Tries == -1; } }
     public double SumReward { get; private set; }
+    public int Tries { get; private set; }
     public double MaxReward { get; private set; }
-    public int Tries { get; private set; }
-
+    public double Value { get { return SumReward / Tries; } }
     public DefaultPolicyActionInfo() {
-      MaxReward = double.NegativeInfinity;
+      MaxReward = double.MinValue;
     }
 
@@ -37 +37 @@
       Console.WriteLine("avg reward {0,5:F2} disabled {1}", SumReward / Tries, Disabled);
     }
+
+    public static Func<DefaultPolicyActionInfo, double> AverageReward {
+      get {
+        return (aInfo) =>
+          aInfo.Tries == 0 ?
+          double.PositiveInfinity :
+          aInfo.SumReward / (double)aInfo.Tries;
+      }
+    }
   }
 }