Learning Dynamic-Objective Policies from a Class of Optimal Trajectories | IEEE Conference Publication | IEEE Xplore