On the Optimality of Batch Policy Optimization Algorithms

Chenjun Xiao * 1 2 Yifan Wu * 3 Tor Littlemore 4 Bo Dai 2 Jincheng Mei 1 2 Lihong Li  5 Csaba Szepesvari 1 4
                       Dale Schuurmans 1 2

             Abstract                 a fixed dataset of previously collected experience, with no
   Batch policy optimization considers leveraging        further environment interaction available. Interest in this
   existing data for policy construction before inter-      problem has ...