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 UCB1Policy : BanditPolicy {
|
---|
10 | private readonly int[] tries;
|
---|
11 | private readonly double[] sumReward;
|
---|
12 | private int totalTries = 0;
|
---|
13 | public UCB1Policy(int numActions)
|
---|
14 | : base(numActions) {
|
---|
15 | this.tries = new int[numActions];
|
---|
16 | this.sumReward = new double[numActions];
|
---|
17 | }
|
---|
18 |
|
---|
19 | public override int SelectAction() {
|
---|
20 | int bestAction = -1;
|
---|
21 | double bestQ = double.NegativeInfinity;
|
---|
22 | foreach (var a in Actions) {
|
---|
23 | if (tries[a] == 0) return a;
|
---|
24 | var q = sumReward[a] / tries[a] + Math.Sqrt((2 * Math.Log(totalTries)) / tries[a]);
|
---|
25 | if (q > bestQ) {
|
---|
26 | bestQ = q;
|
---|
27 | bestAction = a;
|
---|
28 | }
|
---|
29 | }
|
---|
30 | return bestAction;
|
---|
31 | }
|
---|
32 | public override void UpdateReward(int action, double reward) {
|
---|
33 | Debug.Assert(Actions.Contains(action));
|
---|
34 | totalTries++;
|
---|
35 | tries[action]++;
|
---|
36 | sumReward[action] += reward;
|
---|
37 | }
|
---|
38 |
|
---|
39 | public override void DisableAction(int action) {
|
---|
40 | base.DisableAction(action);
|
---|
41 | totalTries -= tries[action];
|
---|
42 | tries[action] = -1;
|
---|
43 | sumReward[action] = 0;
|
---|
44 | }
|
---|
45 |
|
---|
46 | public override void Reset() {
|
---|
47 | base.Reset();
|
---|
48 | totalTries = 0;
|
---|
49 | Array.Clear(tries, 0, tries.Length);
|
---|
50 | Array.Clear(sumReward, 0, sumReward.Length);
|
---|
51 | }
|
---|
52 | }
|
---|
53 | }
|
---|