1 paper
Edoardo David Santi, Gongpu Chen, Deniz Gündüz +1
We consider a Markov decision process (MDP) in which actions prescribed by the controller are executed by a separate actuator, which may behave adversarially. At each time step, th…