source: branches/HeuristicLab.Problems.GrammaticalOptimization/HeuristicLab.Algorithms.Bandits/BanditPolicies/ModelPolicyActionInfo.cs @ 11744

using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.Linq;
using System.Text;
using System.Threading.Tasks;

namespace HeuristicLab.Algorithms.Bandits.BanditPolicies {
  // uses a statistical model to sample and update posterior distribution p(Reward | Data)
  public class ModelPolicyActionInfo : IBanditPolicyActionInfo {
    private readonly IModel model;
    public bool Disabled { get { return Tries == -1; } }
    public double Value { get { return model.SampleExpectedReward(new Random()); } }

    public int Tries { get; private set; }
    public ModelPolicyActionInfo(IModel model) {
      this.model = model;
    }

    public void UpdateReward(double reward) {
      Debug.Assert(!Disabled);
      Tries++;
      model.Update(reward);
    }

    public double SampleExpectedReward(Random random) {
      return model.SampleExpectedReward(random);
    }

    public void Disable() {
      this.Tries = -1;
    }

    public void Reset() {
      Tries = 0;
      model.Reset();
    }

    public void PrintStats() {
      model.PrintStats();
    }

    public override string ToString() {
      return string.Format("disabled {0} model {1}", Disabled, model);
    }
  }
}