Gradient-based policy iteration: an example | IEEE Conference Publication | IEEE Xplore