Dueling Double Deep Q Network

The paper presents a combined task-level reinforcement learning and motion planning framework to solve a multi-class in-rack test tube rearrangement problem. The framework uses reinforcement learning at the task level…