1 paper · 2 filters
Jerome Taupin, Yassir Jedra, Alexandre Proutiere
We investigate the problem of best policy identification in discounted linear Markov Decision Processes in the fixed confidence setting under a generative model. We first derive an…