The multi-armed bandit problem under the mean-variance setting

https://freakonometrics.hypotheses.org/80295