124
click to vote
ICML
2003
IEEE
16 years 4 months ago
2003
IEEE
We investigate methods for planning in a Markov Decision Process where the cost function is chosen by an adversary after we fix our policy. As a running example, we consider a rob...