Research On Quantum System Control Based On Reinforcement Learning

The design of quantum system control has become a key task required to develop powerful quantum information technology,and such research is related to many fields such as quantum information,atomic physics,and physical chemistry.For a class of quantum system control tasks with limited control resources,some reinforcement learning algorithms are used to solve such problems.In addition,from the perspective of improving reinforcement learning algorithms,emotional reinforcement learning algorithm is proposed and applied to solve quantum system control problems.For solving quantum system control problems based on reinforcement learning,the quantum system control task is modeled as a problem that can be solved by reinforcement learning.Meanwhile,three-switch control and Bang-Bang control are defined according to the number of selectable unitary operations.Simulation experiments based on Q-learning,probabilistic q-learning and quantum reinforcement learning realize the task of controlling the quantum system from the initial state to the target state.In terms of improving reinforcement learning algorithms,a reward function based on emotion theory is designed,and emotional reinforcement learning algorithm is proposed.In addition,the comparison experiments between the new algorithm and some traditional reinforcement learning algorithms prove that the new algorithm canaccelerate the learning speed.Finally,emotional reinforcement learning is used to solve the problem of quantum system control,and simulation results show the effectiveness of the new algorithm in solving this problem.