The paper proposes a new methodology to optimize the robot and positioner motions in redundant robotic system for the fiber placement process. It allows user to find time-optimal smooth profiles for the joint variables while taking into account full capacities of the robotic system expressed by the maximum actuated joint velocities and accelerations. In contrast to the previous works, the proposed methodology possesses high computational efficiency and also takes into account the collision constraints. The developed technique is based on conversion of the original continuous problem into a discrete one, where all possible motions of the robot and the positioner are represented as a directed multi-layer graph and the desired time-optimal motions are generated using the dynamic programming that is applied sequentially for the rough and fine search spaces. To adjust the optimization results to the engineering requirements, the obtained trajectories are smoothed using the spline approximation. The advantages of the proposed methodology are confirmed by an application example that deals with a planar fiber placement robotic system.