[11710] | 1 | using System;
|
---|
| 2 | using System.Collections.Generic;
|
---|
[11727] | 3 | using System.Diagnostics;
|
---|
[11710] | 4 | using System.Linq;
|
---|
| 5 | using System.Text;
|
---|
| 6 | using System.Threading.Tasks;
|
---|
| 7 |
|
---|
| 8 | namespace HeuristicLab.Algorithms.Bandits {
|
---|
| 9 | public class UCBNormalPolicy : BanditPolicy {
|
---|
| 10 | private readonly int[] tries;
|
---|
| 11 | private readonly double[] sumReward;
|
---|
| 12 | private readonly double[] sumSqrReward;
|
---|
| 13 | private int totalTries = 0;
|
---|
| 14 | public UCBNormalPolicy(int numActions)
|
---|
| 15 | : base(numActions) {
|
---|
[11727] | 16 | this.tries = new int[numActions];
|
---|
| 17 | this.sumReward = new double[numActions];
|
---|
| 18 | this.sumSqrReward = new double[numActions];
|
---|
[11710] | 19 | }
|
---|
| 20 |
|
---|
| 21 | public override int SelectAction() {
|
---|
[11727] | 22 | Debug.Assert(Actions.Any());
|
---|
[11710] | 23 | int bestAction = -1;
|
---|
| 24 | double bestQ = double.NegativeInfinity;
|
---|
[11727] | 25 | foreach (var a in Actions) {
|
---|
| 26 | if (totalTries == 0 || tries[a] == 0 || tries[a] < Math.Ceiling(8 * Math.Log(totalTries))) return a;
|
---|
| 27 | var avgReward = sumReward[a] / tries[a];
|
---|
[11710] | 28 | var q = avgReward
|
---|
[11727] | 29 | + Math.Sqrt(16 * ((sumSqrReward[a] - tries[a] * Math.Pow(avgReward, 2)) / (tries[a] - 1)) * (Math.Log(totalTries - 1) / tries[a]));
|
---|
[11710] | 30 | if (q > bestQ) {
|
---|
| 31 | bestQ = q;
|
---|
[11727] | 32 | bestAction = a;
|
---|
[11710] | 33 | }
|
---|
| 34 | }
|
---|
| 35 | return bestAction;
|
---|
| 36 | }
|
---|
| 37 | public override void UpdateReward(int action, double reward) {
|
---|
[11727] | 38 | Debug.Assert(Actions.Contains(action));
|
---|
[11710] | 39 | totalTries++;
|
---|
| 40 | tries[action]++;
|
---|
| 41 | sumReward[action] += reward;
|
---|
| 42 | sumSqrReward[action] += reward * reward;
|
---|
| 43 | }
|
---|
[11727] | 44 |
|
---|
| 45 | public override void DisableAction(int action) {
|
---|
| 46 | base.DisableAction(action);
|
---|
| 47 | totalTries -= tries[action];
|
---|
| 48 | tries[action] = -1;
|
---|
| 49 | sumReward[action] = 0;
|
---|
| 50 | sumSqrReward[action] = 0;
|
---|
| 51 | }
|
---|
| 52 |
|
---|
[11710] | 53 | public override void Reset() {
|
---|
[11727] | 54 | base.Reset();
|
---|
[11710] | 55 | totalTries = 0;
|
---|
| 56 | Array.Clear(tries, 0, tries.Length);
|
---|
| 57 | Array.Clear(sumReward, 0, sumReward.Length);
|
---|
| 58 | Array.Clear(sumSqrReward, 0, sumSqrReward.Length);
|
---|
| 59 | }
|
---|
| 60 | }
|
---|
| 61 | }
|
---|