@Article{,
author = "Peter Marbach and John Tsitsiklis",
title = "Approximate Gradient Methods in Policy-Space Optimization of Markov Reward Processes",
publisher = "Kluwer Academic Publishers",
journal = "Discrete Event Dynamic Systems",
volume = "13",
number = "1-2",
month = "January",
year = "2003",
pages = "111–148",
}