[11727] | 1 | using System;
|
---|
| 2 | using System.Collections.Generic;
|
---|
| 3 | using System.Diagnostics;
|
---|
| 4 | using System.Linq;
|
---|
| 5 | using System.Text;
|
---|
| 6 | using System.Threading.Tasks;
|
---|
| 7 |
|
---|
| 8 | namespace HeuristicLab.Algorithms.Bandits {
|
---|
| 9 | public class Exp3Policy : BanditPolicy {
|
---|
| 10 | private readonly Random random;
|
---|
| 11 | private readonly double gamma;
|
---|
| 12 | private readonly double[] w;
|
---|
| 13 | // TODO: debug (very large weights over time)
|
---|
| 14 | public Exp3Policy(Random random, int numActions, double gamma)
|
---|
| 15 | : base(numActions) {
|
---|
| 16 | if (gamma < 0 || gamma > 1) throw new ArgumentException();
|
---|
| 17 | this.random = random;
|
---|
| 18 | this.gamma = gamma;
|
---|
| 19 | this.w = Enumerable.Repeat(1.0, numActions).ToArray();
|
---|
| 20 | }
|
---|
| 21 |
|
---|
| 22 | public override int SelectAction() {
|
---|
| 23 | Debug.Assert(Actions.Any());
|
---|
| 24 | var numActions = Actions.Count();
|
---|
| 25 | var sumW = w.Sum();
|
---|
| 26 | var r = random.NextDouble();
|
---|
| 27 | var sumP = (1 - gamma) * w[0] / sumW + gamma / numActions;
|
---|
| 28 | int i = 0;
|
---|
| 29 | while (r > sumP) {
|
---|
| 30 | i++;
|
---|
| 31 | sumP += (1 - gamma) * w[i] / sumW + gamma / numActions;
|
---|
| 32 | }
|
---|
| 33 | Debug.Assert(i >= 0 && i < numActions);
|
---|
| 34 | return i;
|
---|
| 35 | }
|
---|
| 36 |
|
---|
| 37 | public override void UpdateReward(int action, double reward) {
|
---|
| 38 | Debug.Assert(Actions.Contains(action));
|
---|
| 39 | var numActions = Actions.Count();
|
---|
| 40 | var p = (1 - gamma) * w[action] / w.Sum() + gamma / numActions;
|
---|
| 41 | var estReward = reward / p;
|
---|
| 42 | w[action] = w[action] * Math.Exp(gamma * estReward / numActions);
|
---|
| 43 | }
|
---|
| 44 |
|
---|
| 45 | public override void DisableAction(int action) {
|
---|
| 46 | base.DisableAction(action);
|
---|
| 47 | w[action] = 0;
|
---|
| 48 | }
|
---|
| 49 |
|
---|
| 50 | public override void Reset() {
|
---|
| 51 | base.Reset();
|
---|
| 52 | foreach (var a in Actions) w[a] = 1.0;
|
---|
| 53 | }
|
---|
| 54 | }
|
---|
| 55 | }
|
---|