> Thompson sampling is great. It's intuitive and computationally tractable. The literature is full of other strategies, specifically semi-uniform strategies, but I strongly recommend using Thompson sampling if it works for your problem.
Spot on. For more information about the advantages of Thompson sampling over other approaches, see Why is Posterior Sampling Better than Optimism for Reinforcement Learning? [1] by Osband and Van Roy.
Spot on. For more information about the advantages of Thompson sampling over other approaches, see Why is Posterior Sampling Better than Optimism for Reinforcement Learning? [1] by Osband and Van Roy.
[1] http://proceedings.mlr.press/v70/osband17a/osband17a.pdf